BEGIN:VCALENDAR VERSION:2.0 PRODID:-//CERN//INDICO//EN BEGIN:VEVENT SUMMARY:Seminar in QUANTITATIVE LIFE SCIENCES: "From worm burrows to dinos aur footprints: An ichnological journey into biological behavior" DTSTART;VALUE=DATE-TIME:20141022T090000Z DTEND;VALUE=DATE-TIME:20141022T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-a14346@ictp.it DESCRIPTION:Since Paleolithic times\, humans have tried to read biological behavior through animal and plant traces. Over the centuries\, this art o f survival has evolved into a discipline halfway between Earth Sciences an d Life Sciences: Ichnology. \n\nThis seminar discusses the fundamentals of the study of traces\, presenting robust techniques to decipher the ichnol ogical alphabet – from worm burrows to dinosaur footprints. Special atte ntion will be paid to the application of quantitative techniques such as f ractal geometry\, network theory and artificial life. Fossil and modern en vironments will be explored\, underlining the (paleo)biological significan ce of 600 millions of years of bioturbation in marine and terrestrial ecos ystems. In fact\, burrows\, footprints\, borings\, coprolites are fossil b ehavior and\, as such\, they are a valuable tool in sedimentary geology\, evolutionary biology and paleontology.\n\n//indico.ictp.it/event/a14346/ LOCATION:ICTP Leonardo da Vinci Building Luigi Stasi Seminar Room URL://indico.ictp.it/event/a14346/ END:VEVENT BEGIN:VEVENT SUMMARY:Seminar in QUANTITATIVE LIFE SCIENCES: "Irreversibility and dissip ation in multiple-scale systems" DTSTART;VALUE=DATE-TIME:20141124T100000Z DTEND;VALUE=DATE-TIME:20141124T113000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7309@ictp.it DESCRIPTION:Systems with disparate timescales are ubiquitous in physics an d biology. Efficient modeling dictates to restrict the description to the slower processes and not to fully resolve the faster ones.\n\nThe seminar will discuss this general issue in stochastic systems with multiple timesc ales. In particular\, the presentation will address the behavior of thermo dynamics upon changing the level of resolution.\nStefano Bo will show how for non-equilibrium systems\, under quite general conditions\, properties of \nthe fast dynamics still affect thermodynamics on the slower scales th erefore compromising \na description of the full system in terms of slow v ariables only. \n\nFinally\, two relevant examples will be given concernin g Brownian motion and a simple\nbiochemical network.\n\n//indico.ictp.it/e vent/7309/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/7309/ END:VEVENT BEGIN:VEVENT SUMMARY:Movement Ecology Day DTSTART;VALUE=DATE-TIME:20150129T090000Z DTEND;VALUE=DATE-TIME:20150129T170000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7352@ictp.it DESCRIPTION:Morning session                                                       \n10:00-10:45             Peter McGrath (TWAS) "Movement Ecology: Local Persp ectives involving Insects"\n10:45-11:30            Adrian Tompk ins (ICTP) "Modelling malaria transmission with the dynamical model VECTRI :  seasonal forecasting\, climate change\, land use change and population migration"\n11:30-11:45            Break\n                                     \n11:45-12:30             Antonio Celani (ICTP)  "Scent tracking"Afternoon session                                                                   \n14:00-14:45             Francesca Malfatti (OGS) "Atomic Force Microscopy\,  h igh-speed confocal laser microscopy and NanoSIMS study of marine Synechoco ccus-heterotrophic bacteria interaction reveals  species specificity and potential biogeochemical consequences"\n14:45-15:30            Asja Jelic (ICTP) "Information transfer in moving animal groups: the case of  turning flocks of starlings"\n15:30-15:45            Break \n                                     \n15:45-17:00            Discussion                           \n\n//indico.ictp.it/event/7352/ LOCATION: Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/7352/ END:VEVENT BEGIN:VEVENT SUMMARY:Experimental spinal cord damage in vitro: From molecular mechanism of pharmacological neuroprotection to modelling studies DTSTART;VALUE=DATE-TIME:20150209T100000Z DTEND;VALUE=DATE-TIME:20150209T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7362@ictp.it DESCRIPTION:Full understanding of the complex mechanisms that lead to neur onal death after a spinal cord lesion is important to devise neuroprotecti ve treatments. Amongst the factors proposed to strongly contribute to neur onal loss is the substantially raised concentration of extracellular gluta mate\, the main excitatory neurotransmitter\, evoked by excitotoxic damage . This process is thought to be linked to the onset of the secondary injur y that can amplify the initial damage to distant areas of the spinal cord. \nAt SISSA\, Trieste\, a model that mimics the early pathological process es caused by traumatic injury has recently been developed to follow up its time-dependent evolution. This model allowed us to demonstrate that excit otoxicity-induced by the glutamate primarily destroyed neurons via a proce ss slowly developing over a few hours. However\, the neuroprotection strat egies with several inhibitors of cell death pathways have been disappointi ng in terms of histological and functional outcome\, thus pointing to addi tional death mechanisms. \nDuring my presentation I will discuss: \n1. the main cell targets essential for locomotor patterns upon this pathological condition\; \n2. use of pharmacological drugs to protect against damage\; \n3. the endogenous neuroprogenitor cell proliferation induced by a patho logical condition\; \n4. the early biomarkers to follow the progression\; \n5. importance of cell death spatio-temporal dynamic map to follow topogr aphical and topological extension with an impact on locomotor function. \n 6. The surprising non-linear relation between neuronal death and locomotor network output. Using a multidisciplinary approach\, new data can be obta ined with high translational value for in vivo or clinical studies in the spinal cord field\, and is a subject that will benefit from tailor-made mo delling with predictive value for further scientific inquiry.\n\n//indico. ictp.it/event/7362/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/7362/ END:VEVENT BEGIN:VEVENT SUMMARY:"Modeling cognitive learning and memory" DTSTART;VALUE=DATE-TIME:20150326T150000Z DTEND;VALUE=DATE-TIME:20150326T163000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7422@ictp.it DESCRIPTION:Abstract -  How ideas from physics and mathematics can be imp rovised to make simple models to understand how the brain learns\, memoriz es and does certain other cognitive functions\, will be discussed in simpl e terms. We also attempt to address questions like\, 'why the long-term me mory is robust against trauma'\, and 'how words are stored in mental lexic on'. While we deal with associative memory in some detail we discuss ways whereby other forms of memory\, such as episodal memory\, can be dealt wit h. We also provide a glimpse of how some of these hypotheses can be tested experimentally.\n\n//indico.ictp.it/event/7422/ LOCATION:ICTP Leonardo Building - Lecture Room H URL://indico.ictp.it/event/7422/ END:VEVENT BEGIN:VEVENT SUMMARY:High-level vision in rats: invariant object recognition and its ne uronal substrates DTSTART;VALUE=DATE-TIME:20150423T090000Z DTEND;VALUE=DATE-TIME:20150423T103000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7432@ictp.it DESCRIPTION:The ability to recognize objects despite tremendous variation in their appearance (e.g.\, because of position or size changes) represent s such a formidable computational feat that it is widely assumed to be uni que to primates. Such an assumption has restricted the investigation of it s neuronal underpinnings to primate studies\, which allow only a limited r ange of experimental approaches. In recent years\, the increasingly powerf ul array of optical and molecular tools that has become available in roden ts has spurred a renewed interest for rodent models of visual functions. H owever\, evidence of primate-like visual object processing in rodents is s till very limited and controversial.\n\nIn this seminar\, I will present b ehavioral evidence showing that rats are capable of recognizing visual obj ects in spite of substantial variation in their appearance\, i.e.\, in spi te of changes in size\, position\, illumination\, in-depth rotation and in -plane rotation. I will also show that such a transformation-tolerant (or invariant) recognition is largely accounted by rat ability to spontaneousl y perceive different views/appearances of an object as similar (i.e.\, as instances of the same object). Next\, I will show that rat object recognit ion relies on a shape-based\, multi-featural processing strategy that make s close-to-optimal use of the discriminatory information afforded by the t arget objects across their various appearances. Finally\, I will present s ome preliminary neuronal recordings from rat visual cortical areas that ru n laterally into rat temporal lobe\, showing that the deepest (i.e.\, high er-order) area conveys more information about object identity (also in spi te of variation in object appearance)\, when compared to lower-order (i.e. \, more occipital) areas (including V1). \n\nTaken together\, these findin gs suggest that the rat visual system may serve as a powerful model to stu dy the neuronal substrates of invariant visual object recognition.\n\n//in dico.ictp.it/event/7432/ LOCATION:ICTP Leonardo Building - Oppenheimer Meeting Room URL://indico.ictp.it/event/7432/ END:VEVENT BEGIN:VEVENT SUMMARY:Bacteria as social and multicellular organisms\; the role of quoru m sensing signaling in the synchronization of bacterial behavior DTSTART;VALUE=DATE-TIME:20150424T090000Z DTEND;VALUE=DATE-TIME:20150424T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7433@ictp.it DESCRIPTION:Cell-cell communication in bacteria occurs through small diffu sible chemical signals that affect target gene expression in response to c ell density. This form of intercellular bacterial signaling is known as qu orum sensing and synchronizes metabolic and behavioral activities of a com munity of bacteria. Many of the activities which are regulated involve vir ulence associated factors in bacterial pathogens and symbiotic traits in h ost-beneficial bacteria. This is therefore a mechanism which controls mult icellular behavior in bacteria and deeper understanding promises to shed l ight on the complexities of the host-microbe relationship and may lead to novel therapeutic applications. Bacterial quorum sensing has also introduc ed the aspect of bacteria being social organisms since via quorum sensing regulate the production of secreted ‘public goods’ which are available to all the neighboring bacterial population. During the seminar\, quorum sensing will be introduced and current research lines in ICGEB-Bacteriolog y will be highlighted.\n\n//indico.ictp.it/event/7433/ LOCATION: Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/7433/ END:VEVENT BEGIN:VEVENT SUMMARY:Statistical Mechanics and Neuroscience: the example of place cell modelling DTSTART;VALUE=DATE-TIME:20150508T120000Z DTEND;VALUE=DATE-TIME:20150508T133000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7450@ictp.it DESCRIPTION:Many brain functions are thought to rely on collective behavio urs of large numbers of interacting neurons. This makes statistical mechan ics a valuable toolbox in the attempt to understand neural systems. This t alk with focus on memory and the celebrated "attractor neural network" par adigm. After a general introduction\, I will explain how these ideas can b e applied to the so-called place cells in the hippocampus\, a neural syste m thought to support the memory of space. A model is proposed that can be studied in great details with analytical methods commonly used for disorde red systems. The various collective behaviours it exhibits are thought to be related to recent experimental results.\n\n//indico.ictp.it/event/7450/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/7450/ END:VEVENT BEGIN:VEVENT SUMMARY:Three dimensional growth paradigms: from architecture to invasion DTSTART;VALUE=DATE-TIME:20150513T080000Z DTEND;VALUE=DATE-TIME:20150513T093000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7451@ictp.it DESCRIPTION:One of the hallmarks of cancer is to lose grip on growth contr ol giving rise to a plethora of phenotypes: from disordered growth to inva sion of adjacent tissues. I will discuss two paradigms of normal growth me chanisms gone awry.\nCultured epithelial cells can form cysts\, i.e. perfe ctly ordered spheroidal monolayers delimiting a single lumen. The orderly development of cysts is a process subject to mechanical constraints and de pends on apico-basal polarity. I will present experimental data and modeli ng results on cystogenesis and on the mechanisms that lead to a disrupted architecture. Our results indicate that dynamics has an important role in leading to aberrant phenotypes.\nOftentimes\, once normal architecture is disrupted\, migration of tumor cells to secondary sites becomes more proba ble. I will describe how metastatic cancer cells in three-dimensional envi ronments can collectively migrate as spheroids and fuse into larger and la rger aggregates. I will describe a theoretical model that explains the exp erimental data in terms of 'chemotaxis driven aggregation'. Theoretical hy potheses are supported by experimental data pointing at an autocrine loop at the basis of such behavior.\n\n//indico.ictp.it/event/7451/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/7451/ END:VEVENT BEGIN:VEVENT SUMMARY:Coverage maximization under resource constraints using random walk based strategies DTSTART;VALUE=DATE-TIME:20150518T123000Z DTEND;VALUE=DATE-TIME:20150518T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7456@ictp.it DESCRIPTION:Information dissemination and search in networks are the most frequently performed operations in any large scale distributed system. Max imization of the number of distinctly visited nodes (coverage) under const raints is the prime target of any proposed algorithm for these operations. In order to quickly achieve high coverage\, the existing algorithms tend to proliferate message packets at higher rate. Due to the nondeterministic nature of the algorithms\, the blind proliferations create many redundant visits in the network which reduces the utilization of available resource . For the first time\, we formally capture this existing tradeoff between the available time and available resource to achieve a certain amount of c overage in the network. Using methods from statistical mechanics\, we show ed that coverage can be optimized\, without using any memory\, for a certa in combinations of resource and time. We also provide the algorithm which can achieve that optimal coverage. For the rest of the combinations of res ource and time\, we provide a notion of optimality employing a very little amount memory. We test our hypothesis through extensive computer simulat ions in various kinds of network topologies.\n\n//indico.ictp.it/event/745 6/ LOCATION: Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/7456/ END:VEVENT BEGIN:VEVENT SUMMARY:Statistical physics of opinion formation DTSTART;VALUE=DATE-TIME:20150602T123000Z DTEND;VALUE=DATE-TIME:20150602T143000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7473@ictp.it DESCRIPTION:Application of statistical physics methods to understand and m odel several social phenomena has gained tremendous importance of late. On e of the most studied problem is how opinion forms in a society.\nHighligh ting the connection with statistical physics\, we review the well known mo dels for opinion dynamics and their variants.\n\n//indico.ictp.it/event/74 73/ LOCATION:ICTP Adriatico Guest House - Giambiagi Lecture Hall URL://indico.ictp.it/event/7473/ END:VEVENT BEGIN:VEVENT SUMMARY:"Many body localization and integrals of motion" DTSTART;VALUE=DATE-TIME:20150611T123000Z DTEND;VALUE=DATE-TIME:20150611T131000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7484@ictp.it DESCRIPTION:Interacting quantum systems in a disordered environment can be in a many-body localized phase\, in which transport is suppressed and mem ory of the initial condition is retained in local observables for arbitrar ily long times. In this talk I will introduce the notion of localization f or disordered quantum systems\, and discuss how it implies ergodicity brea king and absence of thermalization. I will then shortly discuss the existe nce of conserved quasi-local quantities for many-body localized systems\, and comment on the explicit construction of these quantities for a particu lar model of interacting fermions on the lattice.\n\n//indico.ictp.it/even t/7484/ LOCATION:ICTP Adriatico Guest House - Giambiagi Lecture Hall URL://indico.ictp.it/event/7484/ END:VEVENT BEGIN:VEVENT SUMMARY: "Statistical scoring functions for protein folding and docking po se recognition" DTSTART;VALUE=DATE-TIME:20150611T131000Z DTEND;VALUE=DATE-TIME:20150611T135000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7485@ictp.it DESCRIPTION:Although it is widely accepted that the tridimensional structu re - and thus\, function - of a protein depends only on its amino acid seq uence\, the issue of finding a correspondence between a particular sequenc e and its stable conformation is still far from being solved.\nA key role in investigating this problem is played by statistical potentials. In part icular\, scoring functions constitute a wide class of algorithms whose aim is to assess the quality of some collection of objects. In this context\, they are employed to evaluate the stability of a number of different puta tive conformations (poses) of a protein whose stable (native) state is unk nown.\nFirst\, the theoretical foundations of the subject will be delineat ed\, then an example will be given by introducing BACH\, a knowledge-based \, statistical scoring function to assess the quality of both folding and docking poses.\n\n//indico.ictp.it/event/7485/ LOCATION:ICTP Adriatico Guest House - Giambiagi Lecture Hall URL://indico.ictp.it/event/7485/ END:VEVENT BEGIN:VEVENT SUMMARY:"Isometric representation of sensory stimulations by the temporal structure of neural activity" DTSTART;VALUE=DATE-TIME:20150611T135000Z DTEND;VALUE=DATE-TIME:20150611T143000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7486@ictp.it DESCRIPTION:Sensory neural activity mediates all perceptions and must thus provide an efficient representation of the environment. In particular\, t he firing dynamics of population of neurons has to be such that the centra l nervous system can recognize the stimulation that elicited it. One cruci al feature of the code is that it must be unambiguous\, meaning that two d ifferent stimulations cannot elicit the same pattern of activity. Such a f eature allows a recognition of the stimulation provided templates exist. H owever\, given the complexity of the manifold of stimulations\, there is a need for a capacity to recognize a continuum of stimulations. It still re mains unclear how the decoding task can be carried to the level of an infi nite number of possible contexts. Here\, using data from human tactile mic roneurography\, we show that\, if readout in an efficient way\, peripheral activity is unambiguous at the utmost point. We show that properly tuned decoders can capture some geometrical regularities of the input space enco ded by primary afferent spiking signals. This provides strong evidence tha t the neural activity can be endowed with a geometry that mimics that of t he tactile stimulations. We argue here that such a faithful geometric orga nisation may be crucial in order to understand human cognitive abilities s uch as learning\, generalizing and categorizing. Our results also suggest that while first-spike latencies are enough to guarantee maximum informati on transmission of tactile stimuli\, entire primary spike trains constitut e a necessary condition to encode isometric input-output mappings\, a like ly basis for generalization in sensory coding. Thus\, spikes gradually sha pe the representation of the stimulation until a faithful one is attained. \n\n//indico.ictp.it/event/7486/ LOCATION:ICTP Adriatico Guest House - Giambiagi Lecture Hall URL://indico.ictp.it/event/7486/ END:VEVENT BEGIN:VEVENT SUMMARY:Extended Plefka Expansion for Stochastic Dynamics DTSTART;VALUE=DATE-TIME:20150616T123000Z DTEND;VALUE=DATE-TIME:20150616T131000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7487@ictp.it DESCRIPTION:We propose an extension of the Plefka expansion\, which is w ell known for the dynamics of discrete spins\, to stochastic differential equations with continuous degrees of freedom and exhibiting generic nonlin earities. The scenario is sufficiently general to allow application to e.g . biochemical networks involved in metabolism and regulation. The main fea ture of our approach is to constrain in the Plefka expansion not just fi rst moments akin to magnetizations\, but also second moments\, specificall y two-time correlations and responses for each degree of freedom. The end result is an effective equation of motion for each single degree of freedo m\, where couplings to other variables appear as a self-coupling to the pa st (i.e. memory term) and a coloured noise. This constitutes a new mean fi eld approximation that should become exact in the thermodynamic limit of a large network\, for suitably long-ranged couplings. For the analytically tractable case of linear dynamics we establish this exactness explicitly b y appeal to spectral methods of Random Matrix Theory\, for Gaussian coupli ngs with arbitrary degree of symmetry.\n\n//indico.ictp.it/event/7487/ LOCATION:ICTP Adriatico Guest House - Giambiagi Lecture Hall URL://indico.ictp.it/event/7487/ END:VEVENT BEGIN:VEVENT SUMMARY:"The cavity method in routing optimization problems on networks" DTSTART;VALUE=DATE-TIME:20150616T131000Z DTEND;VALUE=DATE-TIME:20150616T135000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7488@ictp.it DESCRIPTION:In recent years the cavity method\, or message-passing algorit hm\, has been successfully exploited by statistical physicists to solve co mbinatorial optimization problems such as the K-satisfability (K-SAT). In this talk we introduce this method and outline how it is applied to solve routing problems on communication or traffic networks. In this context the typical situation is where many users want to communicate at the same tim e over a given network. One then wants to optimize the overall routing by minimizing communication path length and the traffic overlap either at nod es or at edges given a set of constraints. We will show how the cavity met hod provides the algorithmic tools to tackle these types of optimization p roblems efficiently.\n\n//indico.ictp.it/event/7488/ LOCATION:ICTP Adriatico Guest House - Giambiagi Lecture Hall URL://indico.ictp.it/event/7488/ END:VEVENT BEGIN:VEVENT SUMMARY:"Identifying relevant positions in proteins by Critical Variable S election" DTSTART;VALUE=DATE-TIME:20150616T135000Z DTEND;VALUE=DATE-TIME:20150616T142000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7489@ictp.it DESCRIPTION:In the last decades development of better inference methods on data samples has been one of the central points of interest for the wide research community focused on complex systems.\nNormally\, the main idea i s that of assuming a model to be the best one describing a system and expl oiting correlation functions – defined in the most suitable way for a gi ven dataset – to find out a sort of hierarchy in the interaction network arising out of a given data sample. The aim is then that of finding out w hether there are and who are the key–players and finally what kind of in teractions are set among them.\nHere we present a natural application of a new method [M. Marsili et al.\, J. Stat. Mech.\, P09003\, 2013]\, renamed Critical Variable Selection\, which does not take directly into account t wo-points correlations but is able to select the best ways for a set of va riables to cluster\, in the particular case maximizing an information amou nt defined on the clustering itself. In protein sequences\, this leads for instance to the selection of a subset of sites turning out to be key–po ints not only along the primary but also along the tertiary structure of t he protein itself.\nThe results obtained by the study of two protein famil ies\, Che-Y receptors and the ion channels\, compared with those turning o ut from the application of Statistical Coupling Analysis\, being a general ization of the Principal Component Analysis [V. Plerou et al.\, Phys. Rev. Lett.\, 83:7\, 1471-74\, 1999\, N. Halabi et al.\, Cell\, 138:4\, 774-86\ , 2009] show interesting overlaps between the two methods\, being the new one also able to identify most of the tightest contacts in the proteins te rtiary structures [F. Morcos et al.\, PNAS\, 108:49\, E12093-301\, 2011].\ n\n//indico.ictp.it/event/7489/ LOCATION:ICTP Adriatico Guest House - Giambiagi Lecture Hall URL://indico.ictp.it/event/7489/ END:VEVENT BEGIN:VEVENT SUMMARY:"Thermodynamics of error corrections" DTSTART;VALUE=DATE-TIME:20150618T090000Z DTEND;VALUE=DATE-TIME:20150618T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7490@ictp.it DESCRIPTION:Biological systems are able to copy information at a finite te mperature with outstanding accuracy. This accuracy is deeply related with the non-equilibrium nature of the copy process. In biochemical reactions\, such as DNA duplication\, different monomers can be distinguished because of their binding energies or via non-equilibrium kinetic mechanisms. I wi ll show how\, in simple copying schemes\, these two discrimination modes a re mutually exclusive and lead to opposite tradeoffs between error\, dissi pation and reaction velocity.  In non-equilibrium multi-step reactions\, such as in kinetic proofreading\, these different modes can be combined to improve the overall accuracy. I will conclude by discussing how the secon d law of thermodynamics can be used to directly relate copying accuracy wi th thermodynamic observables.\n\n//indico.ictp.it/event/7490/ LOCATION:ICTP Adriatico Guest House - Giambiagi Lecture Hall URL://indico.ictp.it/event/7490/ END:VEVENT BEGIN:VEVENT SUMMARY:Nano-mechanics of DNA in severe conditions DTSTART;VALUE=DATE-TIME:20150619T130000Z DTEND;VALUE=DATE-TIME:20150619T143000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7493@ictp.it DESCRIPTION:DNA is perhaps the most important of all biomolecules\, not o nly because it carries the genetic information of biological systems\, but also because of its mechanical flexibility that helps its biological func tion. Thus\, elasticity of DNA plays an important role in its biological function. With a typical length of a few macrometers to several centimete rs\, it is only two nanometers wide. When the molecule is free in aquatic environment and in physiological conditions it behaves like a flexible rod \, but the molecule is usually found in very tight geometries. In eukaryot ic cells\, it tightly fits inside micron-sized cell nuclei\, with many sha rp bends in nano scale. As another example\, it can be encapsulated and fo und in highly confined structures in viral capsids with a size of tens of nanometers. Thus\, understanding the structure of the molecule under such  severe confinements and the way the molecule responds to these geometri cal constraints\, may help to understand DNA function better.\n \nIn th is talk we overview the theoretical models and their results in molecular structure of DNA in nanoscale\, and its mechanical response to external pe rturbations.  \n \n\n//indico.ictp.it/event/7493/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/7493/ END:VEVENT BEGIN:VEVENT SUMMARY:"Universal properties of branching random walks in confined geomet ries" DTSTART;VALUE=DATE-TIME:20150720T090000Z DTEND;VALUE=DATE-TIME:20150720T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7483@ictp.it DESCRIPTION:Precisely quantifying the flow of radiation such as neutrons o r photons through a structural material or a living body represents a long -standing problem in statistical physics. A fundamental question concerns the occupation statistics of the transported particles within the body whe n entering from the outer surface\, ie the distribution of the travelled l ength l and the number of collisions n performed by the stochastic process in the body.\nStochastic radiation transport is modeled by Pearson random walks\, which can be coupled to a birth-death mechanism (neutron multipli cation in fissile materials for instance): a random number of particles em erge from a collision\, which leads to branching particle trajectories. In particular\, branching Pearson random walks with exponentially distribute d lengths stem from assuming that the traversed medium is homogeneous. In this case\, the Markovian nature of the transport process leads to remarka bly simple Cauchy-like formulas\, relating the surface to the volume avera ges of l and n.\nIn many important applications of transport theory\, the hypothesis of uncorrelated scattering centers is however deemed to fail\, which\nthus calls for models based on non-exponential random walks. During this talk\, we will see how such formulas strikingly carry over to the mu ch broader class of branching processes with arbitrary jumps\, and have th us a universal character.\nOur results are key to such technological issue s as the analysis of radiation flow for nuclear reactor design and medical diagnosis and apply more broadly to physical and biological systems with diffusion\, reproduction and death. The proposed formalism may also apply to animal search strategies in the presence of non-exponential displacemen ts.\n\n//indico.ictp.it/event/7483/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/7483/ END:VEVENT BEGIN:VEVENT SUMMARY:"Model-based Assessment of Aspects of User-satisfaction in Bicycle Sharing Systems" DTSTART;VALUE=DATE-TIME:20151015T140000Z DTEND;VALUE=DATE-TIME:20151015T150000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7712@ictp.it DESCRIPTION:Modelling of bike sharing systems as Markov Renewal Process is examined with the aim of capturing and assessing various forms of user (d is)satisfaction. A class of models with minimal assumptions about distribu tions of bicycle parking stations and service requests is developed in whi ch rational commuter behaviour is taken into account. Stochastic time evol ution of these models is studied\, using a generalised version of Gillespi e's exact stochastic integration algorithm that accounts for non-Markovian inter-event times. The model is shown to reproduce quite faithfully the t rip-duration statistics of smaller and larger real bike sharing systems\, such as those in London and Pisa\, including the algebraic `tails' of thes e distributions that are made up of longer cycling trips. The latter are r elated to user's difficulties to find suitable parking places therefore to a potential source of distress. The model also predicts other salient fea tures such as a mode at 10 minutes and crossover behaviour at about 30 min utes. The framework can be extended to include measures either designed to improve\, or anyway to affect\, the user experience with a system\, such as incentives for spontaneous vehicle redistribution. User satisfaction is difficult to assess in real systems because these naturally collect only data about trips that actually\, and thus successfully\, took place giving only partial and biased insight in user satisfaction.\n\n//indico.ictp.it /event/7712/ LOCATION:ICTP\, ex SISSA building Central Area\, 2nd floor URL://indico.ictp.it/event/7712/ END:VEVENT BEGIN:VEVENT SUMMARY:Encoding sequencing data (and time) DTSTART;VALUE=DATE-TIME:20151027T100000Z DTEND;VALUE=DATE-TIME:20151027T113000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7728@ictp.it DESCRIPTION:We will start from a discussion of the algorithmic aspects of production\, analysis\, and storage for short-reads sequencing-data in a g enomic research project. Along this path we will see what are the most est ablished currently used computational tools\, with some emphasis on some a spect in which we believe a deeper interaction between Computer Science an d Biology would be most welcome. With this in mind\, we will conclude with the description of an ongoing study on some intriguing combinatorial and computational problem suggested by the task of clustering data during (can cer-cell sub-population) analysis and classification. Some temporal aspect of this problem will be taken into account.\n\n//indico.ictp.it/event/772 8/ LOCATION:ICTP Central Area\, 2nd floor\, ex-SISSA building URL://indico.ictp.it/event/7728/ END:VEVENT BEGIN:VEVENT SUMMARY:How sensory systems collect evidence over time DTSTART;VALUE=DATE-TIME:20151118T100000Z DTEND;VALUE=DATE-TIME:20151118T113000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7730@ictp.it DESCRIPTION:\n In natural environments\, most events are characterized by stochastic\, “noisy” sensory signals. Noisy stimuli can be reliably id entified only by accumulating evidence over time. We present studies desig ned to elucidate how an uncertain stimulus is integrated over time to prod uce a percept and how that percept is transformed into a decision. Our str ategy is to compare human and rat performance to find fundamental underlyi ng principles. From these principles we have constructed a model that acco unts for all observed data. In behaving rats\, neurons recorded from vario us regions of neocortex express the computations inherent to the model.\n\ n//indico.ictp.it/event/7730/ LOCATION:ICTP Central Area\, 2nd floor\, ex-SISSA building URL://indico.ictp.it/event/7730/ END:VEVENT BEGIN:VEVENT SUMMARY:Fighting antibiotic resistance with antimicrobial peptides (AMPs) DTSTART;VALUE=DATE-TIME:20151119T100000Z DTEND;VALUE=DATE-TIME:20151119T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7736@ictp.it DESCRIPTION:An increasing number of bacterial pathogens are acquiring mult i-drug resistance (MDR) to different classes of antibiotics\, with certain types of Gram-negative bacteria\, such as Acinetobacter baumannii\, Pseud omonas aeruginosa and Klebsiella pneumoniae\, becoming a general health th reat. Therefore new drugs are urgently needed to reduce the risk of danger ous and costly infections\, and antimicrobial peptides (AMPs) represent a valuable candidate as new antimicrobials. AMPs are an important component of the innate immune system\, possessing both antimicrobial and immune-mod ulatory effects\, and are produced among all the living kingdoms. They are small and cationic molecules exerting their activity mainly disrupting ba cterial membranes\, despite not all of them act via a lytic mechanism of a ction. Indeed the class of proline-rich peptides (PR-AMPs) has been shown to act intracellularly and their mode of action has been well characterise d in Escherichia coli\, relying on the inhibition of protein synthesis upo n entry into the cell via the membrane protein SbmA.\nOur efforts are curr ently devoted to evaluate the activity of the PR-AMP Bac7 against clinical ly relevant bacteria\, such as those mentioned above\, in order to broaden the activity spectrum of the peptide. Moreover we are characterising the mechanism of action of Bac7 against Pseudomonas aeruginosa\, showing that the peptide kills this microorganism with a different mode of action with respect to that described for Escherichia coli\, and we are exploiting RNA seq analysis to obtain new insights into the mode action of Bac7 against S almonella typhimurium.\n\n//indico.ictp.it/event/7736/ LOCATION:ICTP Meeting room\, Central Area\, Old SISSA bldg.\, second floor URL://indico.ictp.it/event/7736/ END:VEVENT BEGIN:VEVENT SUMMARY:Hydrogen bonding interactions in hydrophobic environments: from am yloid proteins to charged water clusters DTSTART;VALUE=DATE-TIME:20151126T100000Z DTEND;VALUE=DATE-TIME:20151126T113000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7749@ictp.it DESCRIPTION:"In this talk I will briefly discuss recent work looking at th e importance of hydrogen bonding interactions in hydrophobic environments such as proteins and air-water interfaces: \n\nAmyloid fibrils are peptide /protein aggregates that have been implicated in neurodegenerative disease s such as Alzheimer's. We have recently gotten interested in the anomalous optical properties of the amyloids since they are able to fluoresce even in the absence of aromatic residues/systems consisting of pi-conjugation. Using first principles simulations\, we find that this property is related to proton transfer that occurs long hydrogen bonds in the fibril.\n\nUnde rstanding the molecular origins of the IR spectra is extremely challenging and requires detailed atomistic molecular models that account for both th e quantum mechanical nature of the electrons and nuclei. I will briefly di scuss work on examining the IR spectra of charged water clusters consistin g of either a proton or a hydroxide ion and how this can change depending on their proximity to the air-water interface."\n\n//indico.ictp.it/event/ 7749/ LOCATION:ICTP Central Area\, second floor\, old Sissa bldg. URL://indico.ictp.it/event/7749/ END:VEVENT BEGIN:VEVENT SUMMARY:A Bayesian Model-Selection-Method as a Criterion for Algorithmic R andomness DTSTART;VALUE=DATE-TIME:20160126T130000Z DTEND;VALUE=DATE-TIME:20160126T143000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7772@ictp.it DESCRIPTION:In this talk I discuss a novel methodology using Bayesian infe rence to assess how random a sequence of symbols is. Our test assigns a li kelihood to all possible models that may have generated a given sequence a nd then selects the model with the higher likelihood. In this way we are a ble to infer whether the underlying process was either unbiased or not. In particular\, when theoretically compared with Borel normality criterion f rom algorithmic theory of information and the NIST suite\, our method show s a statistical more robust and stricter criterion.\n\n//indico.ictp.it/ev ent/7772/ LOCATION:ICTP Central Area\, 2nd floor\, ex-SISSA building URL://indico.ictp.it/event/7772/ END:VEVENT BEGIN:VEVENT SUMMARY:May the faculty of language emerge from a glassy phase transition? DTSTART;VALUE=DATE-TIME:20160210T100000Z DTEND;VALUE=DATE-TIME:20160210T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7781@ictp.it DESCRIPTION:In LIMBO we have explored latching dynamics in simplified cort ical networks as a model of spontaneous production\, for example of senten ces.\nLatching\, or jumping from one attractor state to the next\, emerges beyond a critical value of network connectivity\, without any change or e xtra ingredient in organization\, a clue to the possible evolutionary path of a cortex capable of sustaining language.\nIn departing from the simple st model towards the complexity of real languages\, however\, we face the major effect of correlations and of the resulting disorder-dominated dynam ics\, which call for the refinement and application of statistical physics tools to linguistic data.\n\n//indico.ictp.it/event/7781/ LOCATION:ICTP Central Area\, 2nd floor\, ex-SISSA building URL://indico.ictp.it/event/7781/ END:VEVENT BEGIN:VEVENT SUMMARY:The depletion attraction and its role in the physical behaviour of crowded macromolecular systems DTSTART;VALUE=DATE-TIME:20160212T100000Z DTEND;VALUE=DATE-TIME:20160212T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7782@ictp.it DESCRIPTION:In this talk\, I will review the physics of the depletion inte raction\, demonstrating how it can profoundly  influence  the properties of several systems at the interface between bio- and soft-matter physics. This mechanism of “attraction through repulsion”\, quite general in m ulti-component systems where different species intermingle and mutually ex clude their volumes\, will be  proved to be the key determinant of the co mplex phase behaviour of colloid-polymer mixtures [1] as well as the respo nsible for a dramatic change in the conformational and self-entanglement p roperties of a macromolecule in a crowded environment [2]. To this end\, I will also present and discuss some efficient and rigorous coarse-graining methodologies [3] which\, retaining  only the relevant degrees of freedo m\,  allow to explore the large-scale behaviour of these systems while re ducing their complexity from a computational  perspective.\n[1]G. D’Ada mo\, A. Pelissetto and C. Pierleoni\, J. Chem. Phys. 141\, 024902 (2014).\ n[2]G. D’Adamo and C. Micheletti\, Macromolecules 48\, 6337 (2015)\n[3]G . D’Adamo\, A. Pelissetto and C. Pierleoni\, Soft Matter\, 8\, 5151 (201 1).\n\n//indico.ictp.it/event/7782/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7782/ END:VEVENT BEGIN:VEVENT SUMMARY:Gene expression profiling as a tool for the identification of nove l disease signature for neurodegeneration in humans DTSTART;VALUE=DATE-TIME:20160216T130000Z DTEND;VALUE=DATE-TIME:20160216T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7797@ictp.it DESCRIPTION:\n * ACTIVITY CANCELLED *\n Prion diseases or transmissible sp ongiform encephalopathies (TSE) are a class of fatal infectious neurodegen erative disorders whose pathogenesis mechanisms are not fully understood. The diseases manifest as sporadic\, genetic or acquired. So far\, neither specific biomarkers for early diagnosis nor effective therapeutic targets have been identified. The pathological molecular component of the diseases is a misfolded isoform of the prion protein (PrP) denoted as prion.\n Mou nting evidence suggests that in addition to gene coding for the PrP (PRNP) other genes may contribute to the genetic susceptibility of TSE. In this context\, microarray-    based gene expression analyses offer unique to ols to approach neurodegenerative disorders. In particular\, transcriptome profiling can be used to identify altered transcripts in response to path ogens\, and select potential targets for novel therapeutic approaches. Up to date\, a number of studies have been carried out in order to investigat e the gene expression alterations occurring in prion-infected organisms\, but most of them involved animal models such as mice\, sheep and cattle\, which are not closely related to humans. Several studies have been perform ed on non-human primates but none of them have investigated the genomic ou tcome of prion infection. In this presentation\, new sets of data will be presented in an attempt to delineate molecular mechanisms of prion disease s and identify genes that can modulate progression of these fatal maladies in humans.\n\n//indico.ictp.it/event/7797/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7797/ END:VEVENT BEGIN:VEVENT SUMMARY:Dynein transmits polarized actomyosin cortical flows to promote ce ntrosome separation DTSTART;VALUE=DATE-TIME:20160218T100000Z DTEND;VALUE=DATE-TIME:20160218T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7784@ictp.it DESCRIPTION:Centrosomes are the major microtubule organizing centers of an imal cells and play a critical role in mitosis to organize the mitotic spi ndle and orchestrate chromosome segregation. The two centrosomes present a t the onset of mitosis must separate in a timely and accurate fashion alon g the surface of the nucleus to ensure proper bipolar spindle assembly.\nT he microtubule-associated motor dynein plays a pivotal role in centrosome separation\, but the underlying mechanisms remain elusive\, particularly r egarding how dynein coordinates this process in space and time. Where in t he cell do motors act to separate centrosomes? How are forces organized to bring centrosomes in opposite directions? And how robust is this process? \nTo address these questions\, we have dissected centrosome separation in the C.elegans one-cell stage embryo. We have quantitatively measured and m odeled centrosome separation using a combination of 3D fluorescence time-l apse microscopy\, image processing and computational modeling. Our analysi s reveals that centrosome separation is powered by the joint action of dyn ein at the nuclear envelope and at the cell cortex. Strikingly\, we demons trate that dynein at the cell cortex acts as a force-transmitting device t hat harnesses polarized actomyosin cortical flows initiated by the centros omes earlier in the cell cycle. This novel mechanism elegantly couples cel l polarization with centrosome separation\, thus ensuring faithful cell di vision.\n\n//indico.ictp.it/event/7784/ LOCATION:ICTP Common Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7784/ END:VEVENT BEGIN:VEVENT SUMMARY:Dynein transmits polarized actomyosin cortical flows to promote ce ntrosome separation DTSTART;VALUE=DATE-TIME:20160218T100000Z DTEND;VALUE=DATE-TIME:20160218T113000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7804@ictp.it DESCRIPTION:Centrosomes are the major microtubule organizing centers of an imal cells and play a critical role in mitosis to organize the mitotic spi ndle and orchestrate chromosome segregation. The two centrosomes present a t the onset of mitosis must separate in a timely and accurate fashion alon g the surface of the nucleus to ensure proper bipolar spindle assembly.\nT he microtubule-associated motor dynein plays a pivotal role in centrosome separation\, but the underlying mechanisms remain elusive\, particularly r egarding how dynein coordinates this process in space and time. Where in t he cell do motors act to separate centrosomes? How are forces organized to bring centrosomes in opposite directions? And how robust is this process? \nTo address these questions\, we have dissected centrosome separation in the C.elegans one-cell stage embryo. We have quantitatively measured and m odeled centrosome separation using a combination of 3D fluorescence time-l apse microscopy\, image processing and computational modeling. Our analysi s reveals that centrosome separation is powered by the joint action of dyn ein at the nuclear envelope and at the cell cortex. Strikingly\, we demons trate that dynein at the cell cortex acts as a force-transmitting device t hat harnesses polarized actomyosin cortical flows initiated by the centros omes earlier in the cell cycle. This novel mechanism elegantly couples cel l polarization with centrosome separation\, thus ensuring faithful cell di vision.\n\n//indico.ictp.it/event/7804/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7804/ END:VEVENT BEGIN:VEVENT SUMMARY:Statistical Physics of Molecular Evolution DTSTART;VALUE=DATE-TIME:20160329T120000Z DTEND;VALUE=DATE-TIME:20160329T133000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7837@ictp.it DESCRIPTION:Molecular phenotypes\, such as gene expression levels or prote in binding affinities\, are increasingly accessible to quantitative measur ements. Such phenotypes measure molecular functions and are important targ ets of natural selection. However\, the map between encoding DNA sequences and molecular phenotypes is often too difficult to quantify. This lack of knowledge raises an obvious question: which evolutionary properties of a phenotype are universal\, that is\, independent of the molecular details? In this talk\, I will show that universality is an emerging property of co mplex phenotypes\, which are encoded by multiple genomic loci. I will intr oduce a non-equilibrium framework for adaptive dynamics of such phenotypes in time-dependent environments\, and between co-evolving populations. In time-dependent environments\, changes in the environment drive the evolut ion of the species\, but not vice versa. As an example\, I will present st rong evidence that adaptation dominates the evolution of gene expression l evels in Drosophila. Co-evolving populations reciprocally affect the fitne ss of each other\, acting as time-dependent environments with feedback. As an example\, I will show evidence of co-adaptation between interacting ce llular populations of HIV viruses and the antibody repertoire of a patient over the course of an infection. In particular\, I will discuss the condi tions for emergence of highly potent broadly neutralizing antibodies\, whi ch are now recognized as critical for designing an effective vaccine again st HIV.\n\n//indico.ictp.it/event/7837/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7837/ END:VEVENT BEGIN:VEVENT SUMMARY:Problems in physics\, machine learning\, and computational neurosc ience with many scales of length DTSTART;VALUE=DATE-TIME:20160331T140000Z DTEND;VALUE=DATE-TIME:20160331T153000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7833@ictp.it DESCRIPTION:\nNatural images are scale invariant\, with structures that sp an a vast range of length scales – from the smallest within a few dozen pixels to large structures the size of the image itself. In this talk\, I will present a novel statistical mechanical model of natural images rooted in the theory of critical phenomena that sheds light on this hierarchy of length scales\, paving the road for a (nonequilibrium) statistical model of natural images with few parameters. Studies of natural image statistics were first done in neuroscience but have since become of immense interest in machine learning due to the recent advances in deep learning. My resea rch on natural images fits into a bigger picture: the general and common p roblem of "many scales of length" in physics\, machine learning and neural coding.\n\n//indico.ictp.it/event/7833/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7833/ END:VEVENT BEGIN:VEVENT SUMMARY:Linking protein allocation and bacterial phenotypes DTSTART;VALUE=DATE-TIME:20160405T120000Z DTEND;VALUE=DATE-TIME:20160405T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7829@ictp.it DESCRIPTION:Abstract:\nProtein synthesis comes with large costs for the ce lls. Indeed\, it is often tightly regulated in order for the cells to adap t to the surrounding environment. I will discuss different aspects of the relationship between proteome composition and metabolism using different t heoretical tools\, ranging from flux balance analysis\, to coarse grained proteome kinetics\, to kinetic modelling of single reactions.\n\n//indico. ictp.it/event/7829/ LOCATION:ICTP Common Area\, Second floor\, Ex SISSA building\, URL://indico.ictp.it/event/7829/ END:VEVENT BEGIN:VEVENT SUMMARY:Gene transcription dynamics in the early drosophila embryo DTSTART;VALUE=DATE-TIME:20160408T090000Z DTEND;VALUE=DATE-TIME:20160408T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7841@ictp.it DESCRIPTION:Proper development of the adult fly relies on the correct spat ial assignment of cell types during fly development. The different identit ies of cells are determined by sequentially expressing particular subsets of genes in different parts of the embryo.\nBicoid activated Hunchback is essential for Antero-Posterior patterning (among others).\nThe precision o f the timing and spatial expression of the hunchback gene is regulated at the level of gene expression. This process involves the assembly of the tr anscription machinery and depends on the concentrations of the maternal gr adients.\nUsing recent live imaging experiments that span several cell cyc les we can access information on the dynamics of transcription initiation and gain insight into the dynamics of transcription.\nWe develop an analys is approach based on a tailor made time dependent autocorrelation function that overcomes all experimental and computational problems to quantify th e dynamics of transcription initiation. Both eukaryotic and prokaryotic tr anscription can involve more than just switching between an active and ina ctive state\, so we extend our analysis to increasing complexity that allo ws us to infer the effective number of transcription initiation steps and the rates for progressing through them.\nWe first discuss the accuracy and potential of our method on simulated data and then analyse live MS2-MCP t races from developing fruit fly embryos. We identify the bursty nature of transcription initiation from the hunchback promoter. We show that in the anterior region where Bicoid levels saturate the promoter transcription dy namics is extremely reproducible from cell cycle 11 to 13. We give a new m ethod for computing the limiting accuracy of the antero-posterior boundary sharpness using only the parameters of our model.\n \n\n//indico.ictp.it /event/7841/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7841/ END:VEVENT BEGIN:VEVENT SUMMARY:Coexistence in large ecosystems: from structure to function DTSTART;VALUE=DATE-TIME:20160412T143000Z DTEND;VALUE=DATE-TIME:20160412T153000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7838@ictp.it DESCRIPTION:\nBiological systems are composed of many interacting componen ts - in ecological communities\, populations of different species. Though the first models for interacting populations date to almost a century ago\ , the role of species interactions in regulating the response to perturbat ions\, and in shaping community composition\, is yet not well understood. Models for ecological communities have so far concentrated on small system s. Here I show that\, on the other hand\, large ecological communities dis play simple and consistent patterns that can be understood as the result o f elementary stochastic processes. Using random matrix techniques\, I stud y how the properties of the interaction networks influence the stability o f ecological communities and the persistence of the populations. I conclud e with a list of challenges that need to be overcome to have a more comple te understanding of statical and dynamical properties of large ecosystems. \n\n//indico.ictp.it/event/7838/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7838/ END:VEVENT BEGIN:VEVENT SUMMARY:Self-regulation and the stability of large biological networks DTSTART;VALUE=DATE-TIME:20160419T120000Z DTEND;VALUE=DATE-TIME:20160419T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7852@ictp.it DESCRIPTION:Biological networks are many and varied\, including food webs describing who eats whom in an ecosystem\, gene regulation networks in whi ch genes are connected through suppression and enhancement\, or mutualisti c networks in which plants are connected to their pollinators. What they h ave in common is that they are all large (containing thousands of nodes) a nd possess definite structure\, deviating considerably from simple random graphs. In these networks the nodes interact with one another\, but also w ith themselves: when the nodes represent species\, a negative self-effect (self-regulation) generally arises from intraspecific competition for limi ted resources. One important question concerns the role of self-effects in stabilizing network dynamics. While it is easy to construct networks whic h can be stabilized by only a handful of nodes exhibiting self-regulation\ , their structure is highly special\, and quite different from what is obs erved empirically. Here I show that\, in general\, the vast majority of no des must exhibit substantial self-regulation in order to stabilize dynamic s. The argument holds for both random and highly structured empirical netw orks. I show that in the case of competitive networks the arrangement of s elf-interactions leading to the most/least stabilized configurations can b e justified analytically\, and are confirmed by thorough simulations.\n\n/ /indico.ictp.it/event/7852/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7852/ END:VEVENT BEGIN:VEVENT SUMMARY:Temporal organization of cellular self-replication DTSTART;VALUE=DATE-TIME:20160426T120000Z DTEND;VALUE=DATE-TIME:20160426T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7840@ictp.it DESCRIPTION:What determines the growth rate of a cell? Is there a common l aw that describes the connection between size at birth and growth rate for all balanced-growing cells? How exponential growth at all possible? In th is talk I will offer a novel systems analysis perspective on these biologi cal questions.\nI will explain two generic methods to accelerate a complex assembly process beyond its critical path duration - the duration of the longest serial process that is bound to occur. I will then apply these ide as to the process of cellular self-replication and derive a novel type of growth law relating\nthe doubling rate of a cell with the number of progen ies concurrently under production\, and the critical path duration.\n\n//i ndico.ictp.it/event/7840/ LOCATION:ICTP Central Area\, second floor\, old SISSA building URL://indico.ictp.it/event/7840/ END:VEVENT BEGIN:VEVENT SUMMARY:New tricks for old problems: using maximum entropy models to study hippocampal memory DTSTART;VALUE=DATE-TIME:20160503T120000Z DTEND;VALUE=DATE-TIME:20160503T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7863@ictp.it DESCRIPTION:Understanding how memory and learning happen at the level of n eural circuits is one of the fundamental questions in neuroscience\, with broad implications ranging from education\, to neuroprosthetics or brain-m achine interfaces. The contribution of the hippocampus to the formation o f new memories and memory recall has been investigated extensively\, both experimentally and in theoretical models. Nonetheless\, putting together the pieces of the memory puzzle remains a challenge. \n \nDuring the talk I’ll use a two examples from my recent work to illustrate how ide as from machine learning and statistical physics\, in particular maximum e ntropy models\, can provide novel insights into the contributions of the h ippocampus to learning and memory. First\, I’ll discuss a new approach f or characterising the statistical structure of neural activity and it’ s changes during learning.  Second\, I’ll introduce a top-down framew ork for studying memory recall at the level of neural circuits\, which pr ovides computationally well-founded links across scales\, from single s ynapses\, to the integration of signals in the dendritic arbor\, to circu it dynamics\, to the system's level architecture of the brain and to behav ior.\n\n//indico.ictp.it/event/7863/ LOCATION:ICTP Central Area\, Second floor\, ex SISSA building URL://indico.ictp.it/event/7863/ END:VEVENT BEGIN:VEVENT SUMMARY:Cell Division Control in E. coli DTSTART;VALUE=DATE-TIME:20160506T093000Z DTEND;VALUE=DATE-TIME:20160506T103000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7867@ictp.it DESCRIPTION:\n The coordination of cell growth and division is a long-stan ding problem in biology. In particular\, the mechanisms that ensure size h omeostasis and\, at the same time\, adapt cell growth and size to the envi ronment have been the subject of intense research. However\, the answers w ere traditionally hindered by limited statistics on single cells. Contempo rary experimental techniques overcome this problem\, but this progress mus t be combined with new theoretical tools to approach the data. Focusing on E. coli\, we introduced a quantitative method for estimating the variable s controlling the division rate from dynamic data\, and used it to build a minimal stochastic model of cell growth and division. Combining this meth od with large-scale microscopy experiments\, classic quantitative laws rel ating cell size\, doubling time and growth rate of bacterial populations i n different nutrient conditions can be revisited at the single cell level. The main result is the emergence of a combination of universality and ind ividuality in the growth-division laws of single E. coli cells. These two apparently contrasting behaviors emerge naturally from the condition-depen dent modulation of the division control mechanism\, thus actually represen ting two sides of the same coin. Finally\, the simultaneous observation of cell growth and DNA replication dynamics allowed us to pinpoint replicati on initiation and cell division as the two main "checkpoints" for size con trol. This opens the way to more detailed models of the process\, and to r igorous tests of molecular cell-cycle descriptions.\n\n//indico.ictp.it/ev ent/7867/ LOCATION:ICTP Central Area\, Second Floor\, ex SISSA building URL://indico.ictp.it/event/7867/ END:VEVENT BEGIN:VEVENT SUMMARY:Viral Proteins Inhibition via Computer-Aided Drug Design DTSTART;VALUE=DATE-TIME:20160519T120000Z DTEND;VALUE=DATE-TIME:20160519T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7882@ictp.it DESCRIPTION:Hepatitis C Virus (HCV) is a widely spread viral infection tha t affects millions of people worldwide. Liver cirrhosis and Hepatocellular Carcinoma have very strong link to liver viruses B and C. Sofosbuvir is a newly FDA approved drug (Dec. 2013) that acts against the spread of HCV. It is a Non-Structural 5B (NS5B) RNA dependent RNA polymerase (RdRp) Nucle oside Inhibitors (NI). Sofosbuvir is an anti-HCV drug among different comp ounds (e.g. IDX-184\, R7128\, etc.) that were under clinical trials for ye ars. \n\nThe lecture aims at providing a deep insight on the utilization o f molecular modeling in conjunction with Quantitative Structure–Activity Relationship (QSAR) in studying the interaction between activated drugs a nd 12 amino acids constituting a 5 Å region surrounding the conserved act ive site motif GDD of HCV. These studies among many other studies on MERS CoV and Zika virus proteins would accelerate the rate of antiviral approva l in pharmaceutical research and development.\n\n//indico.ictp.it/event/78 82/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7882/ END:VEVENT BEGIN:VEVENT SUMMARY:Mathematical Modeling of Human Antibody Affinity DTSTART;VALUE=DATE-TIME:20160606T120000Z DTEND;VALUE=DATE-TIME:20160606T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7889@ictp.it DESCRIPTION:Affinity defines the attractive force between antibody (Ab) an d antigen (Ag) and represents a key feature of strong immune response. The study of affinity maturation of antibodies following vaccination or infec tion is a growing field of interest for vaccine discovery. Nevertheless th e affinity characterization of human sera is still precluded due to the co mplexity of polyclonal antibodies\, being the sera a mixture of different Abs clones (mAb) of unknown concentration and affinity.\nWe developed a hi gh-throughput methodology\, based on mathematical modeling of the Ag-Ab ca pture profile in the Gyrolab® miniaturized immunoassay platform\, to defi ne an affinity score and to de-convolute polyclonal sera in monoclonal-lik e sub-populations. The mAb-Ag profile in a hydro-dynamical system can be m odeled as a Lévy process and it is described by an approximated Landau di stribution. The estimation of optimal parameters from the de-convolution o f non-Gaussian mixtures allows to measure affinity and quantity in each Ab sub-population.\nThis approach has been validated on four artificial mixt ures\, each composed by two fully characterized monoclonal antibodies with different affinity and run at different concentrations.  Nevertheless\, capture profiles from certain mixtures of mAbs elude our analysis.\nThe ne xt step is to extend the approach by investigating the microscopic dynamic s in the Gyrolab® experimental set-up and by developing a model to explai n all observed capture profiles in polyclonal sera.\nThis method can be su ccessfully used in human sera from subject immunized with different vaccin es to study the maturation of Ab affinity\, also in relationship with othe r serological measurements\, including cellular responses and correlates o f protection\, and therefore to select the best vaccine formulation.\n\n// indico.ictp.it/event/7889/ LOCATION:ICTP Common Area\, 2nd floor\, ex SISSA building URL://indico.ictp.it/event/7889/ END:VEVENT BEGIN:VEVENT SUMMARY:Periodically Driven DNA: Theory and Simulations DTSTART;VALUE=DATE-TIME:20160607T090000Z DTEND;VALUE=DATE-TIME:20160607T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7894@ictp.it DESCRIPTION:We propose a generic model of driven DNA under the influence o f an oscillatory force of amplitude F and frequency v and show the existen ce of a dynamical transition for a chain of finite length. We find that th e area of the hysteresis loop\, Aloop\, scales with the same exponents as observed in a recent study based on a much more detailed model. However\, towards the true thermodynamic limit\, the high-frequency scaling regime e xtends to lower frequencies for larger chain length L and the system has o nly one scaling Aloop ~ ν-1F2. Expansion of an analytical expression for Aloop obtained for the model system in the low-force regime revealed that there is\na new scaling exponent associated with force Aloop ~ ν-1F2.5\, which has been validated by high-precision numerical calculation. By a com bination of analytical and numerical arguments\, we also deduce that for l arge but finite L\, the exponents are robust and independent of temperatur e and friction coefficient.References:\n1. Periodically driven DNA: Theory and Simulation\nSanjay Kumar\, Ravinder Kumar and Wolfhard Janke Phys. Re v. E \, 93\, 010402 (R)(2016)\n2. Statistical mechanics of DNA unzipping u nder periodic force: Scaling behavior of hysteresis loop\nSanjay Kumar and Garima Mishra Phys. Rev. Lett. 110\, 258102\n(2013)\n\n//indico.ictp.it/e vent/7894/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7894/ END:VEVENT BEGIN:VEVENT SUMMARY:Protein Interactions: Integrating Computational Methods and Experi mental Data for Understanding the Binding Specificity DTSTART;VALUE=DATE-TIME:20160608T090000Z DTEND;VALUE=DATE-TIME:20160608T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7895@ictp.it DESCRIPTION:\n Protein-protein interactions play crucial roles in many bio logical processes and responsible for smooth functioning of the machinery in living organisms. Predicting the binding affinity of protein-protein co mplexes and understanding the recognition mechanism are challenging proble ms in computational and molecular biology (1\,2). We have developed a gene ralized energy based approach for identifying the binding site residues an d interacting pairs in all types of protein complexes. We observed that th e residues with charged and aromatic side chains are important for binding in protein-protein complexes. These residues influence to form cation–p \, electrostatic and aromatic interactions. Our observations have been ver ified with the experimental binding specificity of protein-protein complex es and found good agreement with experiments (3). Further\, we have develo ped algorithms for discriminating protein-protein complexes based on their binding affinities (4) and predicting the binding affinity (5). We sugges t that our method would serve as an effective tool for identifying the int eracting partners in protein-protein interaction networks and human-pathog en interactions based on the strength of interactions (6).\nRecently\, we have developed a protocol to identify the novel inhibitors for cYes kinase \, which is a target for colorectal cancer using homology modeling\, docki ng and virtual screening (7). The lead compounds have been verified with e xperiments\, which show the inhibition rate of 30-70% and the IC50 in the range of 5-30 µM. The salient features of the results will be discussed.R eferences\n1. M.M. Gromiha (2010) Protein Bioinformatics\, Elsevier Publis hers/Academic Press\n2. M.M. Gromiha and K. Yugandhar (2016) Prog. Biophys . Mol. Biol. (in press)\n3. M.M. Gromiha\, K. Yokota and K Fukui (2009) Mo l. Biosystems 5: 1779-1786.\n4. K. Yugandhar and M.M. Gromiha (2014) Prote ins. 82: 2088-96.\n5. K. Yugandhar and M.M. Gromiha (2014) Bioinformatics. 30: 3583-9.\n6. K. Yugandhar and M.M. Gromiha (2016) Curr. Prot. Pept. Sci . 17\, 72-81\n7. S. Chiba\, K. Ikeda\, T. Ishida\, M.M Gromiha et al. (201 5) Sci. Rep. 5\, 17209\n\n//indico.ictp.it/event/7895/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7895/ END:VEVENT BEGIN:VEVENT SUMMARY:How To Make Good Decisions - BASIC NOTIONS SEMINAR SERIES 2016 DTSTART;VALUE=DATE-TIME:20160624T150000Z DTEND;VALUE=DATE-TIME:20160624T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7891@ictp.it DESCRIPTION:Why is decision-making so difficult? There are three main reas ons: first\, our limited ability in computing optimal strategies\; second\ , the incomplete knowledge of the world around us\; third\, the inaccuracy of the predictions about the consequences of our actions. Under this seve re conditions\, is it still possible to act optimally?\nWith the aid of ex amples from biology\, medicine and computer science\, I will discuss how t he theory of Markov Decision Processes and Reinforcement Learning provides an answer to this question.\n\n//indico.ictp.it/event/7891/ LOCATION: URL://indico.ictp.it/event/7891/ END:VEVENT BEGIN:VEVENT SUMMARY:Modeling vehicular traffic in the Philippines DTSTART;VALUE=DATE-TIME:20160630T090000Z DTEND;VALUE=DATE-TIME:20160630T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7912@ictp.it DESCRIPTION:\n Traffic systems modeling has largely been an engineering pr oblem and entire software packages built around the problem currently exis t. Yet some assumptions may be worth questioning. In this talk\, I will pr ovide an overview of some unique features of vehicular traffic in the Phil ippines. These features\, which range from driving behavior to management methods\, provide an interesting test bed for traffic models with adaptive agents.\n\n//indico.ictp.it/event/7912/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7912/ END:VEVENT BEGIN:VEVENT SUMMARY:RNAs competing for microRNAs mutually influence their fluctuations in a microRNA-dependent manner DTSTART;VALUE=DATE-TIME:20160712T090000Z DTEND;VALUE=DATE-TIME:20160712T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7909@ictp.it DESCRIPTION:\n Distinct RNA species may compete for binding microRNAs\, a class of small non-coding transcripts involved in post-transcriptional dow nregulation. This competition creates an indirect interaction between micr oRNA targets that behave as microRNA sponges and eventually influence each other expression levels. Theoretical predictions suggest that not only th e mean target expression levels but also the fluctuations around the means are coupled through microRNAs\, potentially resulting in even more striki ng effects on a broad range of cellular processes. With an experimental de sign based on two bidirectional plasmids and flow cytometry measurements o f cotransfected mammalian cells\, we validated a stochastic gene interacti on model that describes how mRNAs can influence each other’s fluctuation s in a microRNA-dependent manner in single cells.\n\n//indico.ictp.it/even t/7909/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7909/ END:VEVENT BEGIN:VEVENT SUMMARY:Structure-Dynamics Interplay in Directed Complex Networks DTSTART;VALUE=DATE-TIME:20160727T120000Z DTEND;VALUE=DATE-TIME:20160727T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7919@ictp.it DESCRIPTION:Synchronization in large populations of interacting elements h as been observed in different disciplines such as system biology\, neurosc ience\, computer science\, and engineering. During the past decades\, many efforts have been devoted to understand the synchronization phenomena fro m the complex networks point of view. Generally\, the problem of optimizin g synchronization is challenging due to the fact that macroscopic dynamics depend both on the topology of the interaction network and dynamics of th e\n\n individual systems. In this talk  we study the effects of direction ality on network synchronization by assigning directions to the links of a n undirected network. We show that\, the way we assign directions to the l inks to enhance the synchronization depends on both the undirected topolog y of the graph and intrinsic frequencies of the nodes.\n\n//indico.ictp.it /event/7919/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7919/ END:VEVENT BEGIN:VEVENT SUMMARY:Approximation solutions for collective behaviour of self propelled particles DTSTART;VALUE=DATE-TIME:20160727T130000Z DTEND;VALUE=DATE-TIME:20160727T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-7927@ictp.it DESCRIPTION:It is well known that macroscopic behaviour of the microscopic interacting particles obeys hydrodynamics equations. However\, it is a re cent interest to find collective behaviour of active matters\,  specially to describe hydrodynamics of the matter composed of  self propelled part icles. Obtaining  hydrodynamics equations from microscopic interaction ru les in active matter requires approximation. We are going to compare some   approaches.\n\n//indico.ictp.it/event/7927/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/7927/ END:VEVENT BEGIN:VEVENT SUMMARY:Optimal conditions of external medium and intracellular environmen t for efficient chemotactic performance in E. coli DTSTART;VALUE=DATE-TIME:20160930T130000Z DTEND;VALUE=DATE-TIME:20160930T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8023@ictp.it DESCRIPTION:Chemotaxis is the directed motion of organisms in response to the chemical gradient. When E. coli bacterium is placed in a medium with a concentration gradient of nutrient\, it moves up the gradient and accumul ates in regions with higher nutrient concentration. The efficient chemotac tic performance is characterised by the ability to find the favorable regi on quickly and to localize in the favorable region at large times. We inve stigate how this efficiency depends on the external environment and the in ternal biochemical pathway of the E.coli cell. When the cell is in a mediu m where the nutrient is diffusing and the form of nutrient profile is Gaus sian\, we find that there exists an optimal width of the profile for which the search time becomes minimum. In the case when the nutrient diffusion and cell movement occurs over comparable time-scales\, there exists an opt imum value of the nutrient diffusivity for which the search time becomes m inimum. The simulation results in a phenotype\n\n model agree well with ou r analytical calculations in a related coarse-grained model where the bact erium behaves like a random walker with position dependent drift velocity and diffusion coefficient. In a single cell chemotaxis the internal bioche mical pathway involves noise due to the fluctuations present in the number of different molecules taking part in the reactions. This noise plays a v ery crucial role on the chemotactic performance of E. coli bacterium. We f ind that in the long time limit the  localization and the uptake\, i.e. t he total nutrient intercepted along its trajectories\, becomes maximum for an optimum value of the noise strength. We discuss a simple mechanism to explain this effect.\n\n//indico.ictp.it/event/8023/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8023/ END:VEVENT BEGIN:VEVENT SUMMARY:Statistical Learning and the Analysis of Stochastic Models DTSTART;VALUE=DATE-TIME:20161018T090000Z DTEND;VALUE=DATE-TIME:20161018T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8051@ictp.it DESCRIPTION:In this seminar I will present recent work on the use of stati stical learning tools for the analysis of stochastic dynamic model of comp lex systems\, biological ones in primis. Statistical learning offers power ful tools to manage uncertainty\, particularly in model parameters. Statis tical learning methods\, combined with model simulations\, can be used to estimate efficiently the probability of observing a given behaviour as a f unction of model parameters. This can be turned into a jackknife method to perform several tasks: optimise likelihoods efficiently\, doing parameter synthesis\, design\, and control. At the same time\, statistical learning provides novel ways to look into model abstraction\, something we called statistical correction maps.  In this talk\, I will give an overview of t hese ideas\, focussing on some aspects with more detail. \n\n//indico.ict p.it/event/8051/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8051/ END:VEVENT BEGIN:VEVENT SUMMARY:Distinct timescales of population coding across cortex DTSTART;VALUE=DATE-TIME:20161108T100000Z DTEND;VALUE=DATE-TIME:20161108T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8056@ictp.it DESCRIPTION:Abstract:\nThe cortex represents information across widely var ying timescales. For instance\, sensory cortex encodes stimuli that fluctu ate over milliseconds\, whereas in association cortex behavioural choices can require the maintenance of information over seconds. It is poorly unde rstood how the cortex achieves such diverse timescales of information codi ng. While recent work has identified different timescales in features intr insic to individual neurons\, the timescales of information coding in popu lations of neurons have not been studied\, and population codes have not b een compared in depth across cortical regions. In this talk\, I will compa re coding for sensory stimuli and behavioural choices in auditory cortex ( AC) and posterior parietal cortex (PPC) in the mouse brain\, as the animal performs a sound localisation task. I will show that population codes are essential to achieve long and diverse coding timescales\, and that the st atistical structure of the codes differs between sensory and association c ortices.\nAmong PPC neurons\, correlations that are not explained by task events ("functional coupling") extend over long time lags\, and contribute to a long timescale population code characterised by consistent represent ations of choice lasting over two seconds. In contrast\, coupling among AC neurons is weak\, shorter-lived\, and results in moment-to-moment fluctua tions in stimulus and choice information. This suggests that population co upling is a variable property that affects the timescale of information co ding: relatively uncoupled activity in sensory cortex is key for signals t hat change rapidly to code temporally variable stimuli\, whereas highly co upled activity in association cortex appears critical to form a consistent signal from which temporally integrated information can be read out insta ntaneously to drive behaviour.\n\n//indico.ictp.it/event/8056/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8056/ END:VEVENT BEGIN:VEVENT SUMMARY:Computational and analytical approaches to modeling solid tumors DTSTART;VALUE=DATE-TIME:20161116T133000Z DTEND;VALUE=DATE-TIME:20161116T143000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8057@ictp.it DESCRIPTION:Abstract:\nIn the first part of this seminar I shall give a sh ort overview of the work of my group in the field of tumor modeling. In th e second part\, I shall discuss the challenges in this field and how we ar e dealing with them\, with special emphasis on the microenvironment of sol id tumors.\n\n//indico.ictp.it/event/8057/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8057/ END:VEVENT BEGIN:VEVENT SUMMARY:Cox-process representation and inference for stochastic reaction-d iffusion processes DTSTART;VALUE=DATE-TIME:20161117T100000Z DTEND;VALUE=DATE-TIME:20161117T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8058@ictp.it DESCRIPTION:Abstract:\n \nComplex behaviour in many systems arises from t he stochastic interactions of spatially distributed particles or agents. S tochastic reaction-diffusion processes are widely used to model such behav iour in disciplines ranging from biology to the social sciences\, yet they are notoriously difficult to simulate and calibrate to observational data . On the other hand\, spatio-temporal point processes offer several comput ational advantages from the statistical perspective\, and can be coupled t o dynamics via an evolving intensity f! ield. In this talk\, I will discus s how joint time marginals of a stochastic reaction-diffusion process can be approximated in a mean-field sense by a spatio-temporal Cox process. Th e resulting approximation allows us to naturally define an approximate lik elihood\, which can be optimised with respect to the kinetic parameters of the model. We show on several examples from systems biology and epidemiol ogy that the method yields consistently accurate parameter estimates\, and can be used effectively for model selection.\n \nRef: Schnoerr\, Grima a nd Sanguinetti\, Nature Communications 7\, 2016 also http://arxiv.org/abs /1601.01972\n Guido Sanguinetti is a Reader in Machine Learning at the Sc hool of Informatics\, University of Edinburgh. After a degree in Physics ( Genova) and a PhD in Mathematics (Oxford)\, his interests shifted towards machine learning and computational biology\, with a particular focus on de vising statistical models for the dynamics of gene regulation. He has publ ished over 60 articles in leading specialist and interdisciplinary journal s\, including Science\, PNAS\, Nature Communications\, Bioinformatics\, an d JASA. He holds an ERC starting grant and is the recipient of the 2012 PN AS Cozzarelli Prize in Engineering and Applied Science.\n\n//indico.ictp.i t/event/8058/ LOCATION:ICTP Common area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8058/ END:VEVENT BEGIN:VEVENT SUMMARY:Rapid adaptation and the predictability of evolution DTSTART;VALUE=DATE-TIME:20161219T100000Z DTEND;VALUE=DATE-TIME:20161219T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8089@ictp.it DESCRIPTION:Evolution is simple if adaptive mutations appear one at a time . However\, in large microbial populations many mutations arise simultaneo usly resulting in a complex dynamics of competing variants. I will discuss recent insight into universal properties of such rapidly adapting populat ions and compare model predictions to whole genome deep sequencing data of HIV-1 populations at many consecutive time points. Genetic diversity data can further be used to infer fitness of individuals in a population sampl e and predict successful genotypes. We validate these prediction using his torical influenza virus sequence data. Successful predictions of the compo sition of future influenza virus population could guide strain selection f or seasonal influenza vaccines.\n\n//indico.ictp.it/event/8089/ LOCATION:ICTP\, Central Area 2nd floor\, old SISSA building URL://indico.ictp.it/event/8089/ END:VEVENT BEGIN:VEVENT SUMMARY:Collective behavior in systems of self-propelled particles and the computational approach DTSTART;VALUE=DATE-TIME:20161220T100000Z DTEND;VALUE=DATE-TIME:20161220T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8108@ictp.it DESCRIPTION:Many natural and artificial systems consist of agents that hav e ability to show persistent motion. Flock of birds\, school of fish\, hum an crowd\, bacterial colonies are few of the examples. Agents in such syst em are called `Self-propelled Particles' (SPPs). Another remarkable proper ty of the agents in such systems is the ability to exhibit collective beha vior. The beautiful pattern formation in flock of Starling birds or school of fish is just one of the many outcomes of the collective behavior[1].\n There are numerous computational models to understand the collective behav ior in such systems. These models aim to understand the formation of such patterns\, the mechanism behind the collective behavior. One of the notabl e model is given by Vicsek et al. now commonly known as the `Vicsek model' . This minimal model assumes that the agent in the system try to behave as it's neighbors do[2]. With this simple but yet effective assumption this model showed some of the prominent features of the collective behavior.\nI n this session we shall discuss about the systems with SPPs\, the Vicsek m odel and one of the modification to the Vicsek model motivated by the fact that biological agents such as humans\, birds\, fish etc. have restricted view-angle and hence such agents interact anisotropically with its neighb ors if the interaction is via visual clues. We shall discuss the effects o n the collective behavior of SPPs due to this modification within the scop e of the Vicsek model.\n \nRef.:\n[1] T. Vicsek and A. Zafeiris\, Phys. R ep. 517\, 71 (2012).
\n[2] T. Vicsek\, A. Czir ok\, E. Ben-Jacob\, I. Co hen\, and O. Shochet\, Phys. Rev. Lett. 75\, 1226 (1995).\n\n//indico.ictp .it/event/8108/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8108/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Hierarchical Bayesian modelling of volatile environme nts DTSTART;VALUE=DATE-TIME:20170308T100000Z DTEND;VALUE=DATE-TIME:20170308T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8181@ictp.it DESCRIPTION:Abstract:\nWhile Bayesian inference on static quantities is wi despread and has become standard in many contexts\, Bayesian inference in volatile environments presents an additional set of challenges. These aris e when inferring the continually changing states of an agent’s environme nt. To make accurate predictions about future observations\, such an agent needs to model how states change. I will set out these challenges in deta il and introduce recently developed methods for meeting them. In particula r\, I will go into the Hierarchical Gaussian Filter (HGF)\, a generic hier archical model of inference on volatile and uncertain states. In this cont ext\, I will also give examples of neuroscientific studies where the HGF w as used.\n\n//indico.ictp.it/event/8181/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8181/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Higher-order interactions stabilize the dynamics of e cological communities DTSTART;VALUE=DATE-TIME:20170502T150000Z DTEND;VALUE=DATE-TIME:20170502T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8222@ictp.it DESCRIPTION:\n The difficulty of reconciling the staggering biodiversity f ound in tropical rainforests with classical theories of resource partition ing has led ecologists to explore neutral theories of coexistence\, in whi ch all species are assumed to have the same physiological parameters\, and variations in species abundance arise from stochastic fluctuations. Simpl e neutral models have led to much progress\, for example to the investigat ion of spatial and temporal ecological patterns and to the formulation of sampling models that allow contrasting theory and data. However\, the high sensitivity of neutral models to slight perturbations of the parameters a nd the prediction of a strong correlation between a species’ abundance and its age are considered problematic. Here we propose a theory of coexis tence in which all species have different physiological rates\, and intera ct with each other through a network of competitive interactions. We show that our models produce robust coexistence of many species even when param eters are drawn at random. Importantly\, the dynamical stability of our mo dels is due to higher-order interactions — interactions involving more t han two species at a time. The existence of higher-order interactions has been debated in ecology for decades\, but their role in shaping ecological communities is still understudied. Our results show that higher-order int eractions can have dramatic effects on the dynamics of ecological systems.  When set in a stochastic framework\, we recover many results from neutr al theory but improve on the relationship between age and abundance.\n\n //indico.ictp.it/event/8222/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8222/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Bacterial finite-size effects for population expansio n under flow DTSTART;VALUE=DATE-TIME:20170515T123000Z DTEND;VALUE=DATE-TIME:20170515T133000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8223@ictp.it DESCRIPTION:\n For organisms living in a liquid ecosystem\, flow and flow gradients have a dual role as they transport nutrient while\, at the same time\, dispersing the individuals. In absence of flow and under homogeneou s conditions\, the growth of a population towards an empty region is usual ly described by a reaction-diffusion equation. The effect of fluid flow is not yet well understood and the interplay between transport of individual s and growth opens a wide scenario of possible behaviors. In this work\, w e study experimentally the dynamics of non-motile E. coli bacteria colonie s spreading inside rectangular channels\, in PDMS microfluidic devices. By use of a fluorescent microscope we analyze the dynamics of the population density subjected to different co- and counter-flow conditions and shear rates. A simple model incorporating growth\, dispersion and drift of finit e size beads is able to explain the experimental findings. This indicates that models based on the Fisher-Kolmogorov-Petrovsky-Piscounov equation (F KPP) may have to be supplemented with bacterial finite-size effects in ord er to be able to accurately reproduce experimental results for population spatial growth.\n\n//indico.ictp.it/event/8223/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8223/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Digital Epidemiology: Challenges in Data Collection i n Developing Countries DTSTART;VALUE=DATE-TIME:20170516T090000Z DTEND;VALUE=DATE-TIME:20170516T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8226@ictp.it DESCRIPTION:\nRespiratory and other close-contact infectious diseases\, su ch as TB\, measles and pneumonia\, are major killers in much of the develo ping world. Understanding how the diseases spread and for identifying how best to control it can be tackled by modelling the spread diseases. Althou gh central to the models\, few quantitative data are available on relevant contact patterns\, and no study to measure these factors has yet been att empted in developing countries.\nWe have originally exploited device conne ctivity traces from the real world for modelling social network structure. The initial motivation was providing delay tolerant networks among smart phones formed by people. The empirical study of contact networks shares ma ny issues with network-based epidemiology\, and our work has been extended towards understanding the epidemic spread of infectious diseases. Capturi ng human interactions will provide an empirical\, quantitative measurement of social mixing patterns to underpin mathematical models of the spread o f close-contact diseases.\nI will describe remote sensing platform to coll ect human mobility data using RFID sensors\, Raspberry Pis and mobile phon es\, recording proximity\, to gather information on human interactions in rural and urban communities in developing countries.\n\n//indico.ictp.it/e vent/8226/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8226/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Extracting information from simple statistical laws i n complex component systems DTSTART;VALUE=DATE-TIME:20170517T130000Z DTEND;VALUE=DATE-TIME:20170517T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8228@ictp.it DESCRIPTION:\n Several complex systems in various fields can be described as component systems\, i.e. sets of objects (genomes\, books\, or LEGO toy s) composed of elementary components (genes\, words\, or LEGO bricks). Sev eral emerging statistical laws regarding the statistics of components  ca n be empirically observed in such diverse systems. These laws may be the c onsequence of  the underlying architectural constrains\, thus in principl e can provide information about the system properties.\n\n Our work  tack les the general questions of what can be learned from these simple statist ical laws about what laws  are a "universal" property of very different c omponent system and what are instead specific of the system in analysis\, how and if these laws are related to each other\, and what simple stochast ic processes can be used to understand their origins. \n In this presenta tion I will focus on two specific examples. The first example concerns the "U"-shaped distribution of  shared genes across genomes\, which is centr al to the current debate in evolutionary genomics. We show that its charac teristic shape can be obtained by a null model simply based on the empiric al heterogeneity of the component abundances. This implies that the distri bution of shared genes is mainly a statistical consequence of other known system properties. This result shows that to extract the relevant biologic al information it is necessary to build null models. In this way it is pos sible to take into account general emerging features and thus extract the systems specific properties.\n The second example considers the growth of the book/object "vocabulary" (i.e. how many distinct words/components are present) as a function of its "size" (i.e. the total number of components) \,  a law known in linguistics as  Heaps’ law. We focused on how the v ocabulary fluctuations scale with its average value\, showing a non-trivia l and general behavior across different systems. Specifically\, the standa rd deviation grows linearly with the average (Taylor’s law). We have fou nd that the minimal stochastic growth processes that can reproduce this sc aling belong to a class of models that includes  the Chinese Restaurant p rocess. This suggests a general rich-gets-richer mechanism in the innovati on dynamics of thus component systems\, leading to interesting system-spec ific interpretations.\n \n\n//indico.ictp.it/event/8228/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building\, Via Beirut URL://indico.ictp.it/event/8228/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Aggregation of wage and capital income in economy DTSTART;VALUE=DATE-TIME:20170619T120000Z DTEND;VALUE=DATE-TIME:20170619T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8265@ictp.it DESCRIPTION:\n Aggregation of different values in the complex systems is o ne of the major goals of the science of complexity. In economy there are s ome known stylized facts which address relations between different aggrega te values. Within them the relation between aggregate wage income and nati onal income has been of the most interesting ones. Labor share of income h as been sustainable over a century and has been about two third of the nat ional income in the United States. Cobb and Douglass suggested the hypothe sis of the aggregate production function to address the roots of the susta inability. Their hypothesis however has been under huge criticism. In this talk we introduce the theory of production function. We then look over th e theory from the statistical physics point of view. In the following we a ggregate capital in a soup of firms. In the end we present a new proposal for the roots of the sustainability of the labor share of income. \n\n//i ndico.ictp.it/event/8265/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building\, via Beirut URL://indico.ictp.it/event/8265/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Exploration Behavior and Its Shaping Effect on Spatia l Memory DTSTART;VALUE=DATE-TIME:20170621T090000Z DTEND;VALUE=DATE-TIME:20170621T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8258@ictp.it DESCRIPTION:\n When exposed to novel environments\, animals may switch bet ween curious and anxious states. The exploration behaviour is therefore hi ghly dynamic and complex. The build-up of internal representation of the e nvironments is coupled with exploration. In this talk\, I will review rece nt progress on the spatial representations in the medial entorhinal cortex of rodents\, and discuss on-going work on neural network models of spatia l memory and the analysis of free exploration behaviour. \n\n  \n\n//ind ico.ictp.it/event/8258/ LOCATION:ICTP\, Central Area 2nd floor\, old SISSA building\, Via Beirut URL://indico.ictp.it/event/8258/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Mitotic waves in the early embryogenesis of Drosophil a: bistability traded for speed DTSTART;VALUE=DATE-TIME:20170712T090000Z DTEND;VALUE=DATE-TIME:20170712T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8277@ictp.it DESCRIPTION:\n Early embryogenesis of most metazoans is characterized by r apid and synchronous cleavage divisions. In Drosophila embryos\, waves of activity of Cdk1\, the main regulator of the cell cycle\, are responsible for the synchronization of cell divisions. After reviewing the phenomenolo gy\, I shall discuss the dynamics of the Ginzburg–Landau model that capt ures the in vivo dynamics of Cdk1. The main feature of the corresponding p otential is that it features a time-dependent component\, which accounts f or the growing level of Cdk1 activation across the cell cycle. Two distinc t regimes result. The first regime is experimentally observed in mutants t hat alter the regulation of the mitotic switch. In the resulting quasi-adi abatic regime\, waves reflect the classical physical mechanism of invasion by a stable state of a metastable one. Conversely\, the quasi-adiabatic a pproximation is violated in wild-type embryos\, where the observed waves r eflect the sweeping of spatial gradients of the Cdk1 field by the time-dep endent component of the potential. This alternative mechanism leads to wav e-like spreading\, which is faster and depends on the dynamic parameters d ifferently than bistable waves. For instance\, the latter scale as the squ are root of the Cdk1 molecular diffusivity and weakly depend on the amplit ude of the noise\, whilst the former scale as 3/4 and 1/2 powers\, respect ively. Theoretical predictions are supported by experiments that couple qu antitative measurements of Cdk1 and genetic perturbations.\n\n  \n\n//ind ico.ictp.it/event/8277/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8277/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Multi-scale analysis of protein-protein interactions: from residues to phylogeny DTSTART;VALUE=DATE-TIME:20170914T093000Z DTEND;VALUE=DATE-TIME:20170914T103000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8378@ictp.it DESCRIPTION:\n Proteins are an essential part of life and participate in v irtually every process within cells. Proteins rarely act alone\, and in re cent years\, the problem of reconstructing the proteins interactome from t he large database of available proteins sequences has triggered much inter est. Yet\, this network has been shaped by various factors and constraints (chemical interactions between residues\, evolution by mutations...). Aft er an overview of the biological forces behind protein protein interaction s\, I will present a "first principle" statistical analysis\, to infer put ative interactions between proteins based on their sole amino-acids sequen ces.\n\n//indico.ictp.it/event/8378/ LOCATION:ICTP Common Area\, 2nd floor\, SISSA building URL://indico.ictp.it/event/8378/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Modeling vascular tumors DTSTART;VALUE=DATE-TIME:20171009T123000Z DTEND;VALUE=DATE-TIME:20171009T133000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8401@ictp.it DESCRIPTION:\n Tumors appear in various types and affect a huge variety of functional aspects. We focus on the stage when the cluster of cancerous c ells becomes larger than the average diffusion length of the required nutr ients. At this point the tumor starts to modify the surrounding vascular s tructure of healthy tissue which is a complex and not yet understood proce ss. Since the changes in topology and spatial arrangement of blood vessels and capillaries affect the distribution of interstitial fluid\, oxygen\, etc. this process is of special interest in order to understand cancer at malignant scales.\n We use computer simulations to design artificial blood vessel networks followed by modeling vascular tumor growth. During this t alk I will give you a brief overview of our model.\n  \n\n//indico.ictp.i t/event/8401/ LOCATION:ICTP Common Area\, second floor\, old SISSA building\, via Beirut URL://indico.ictp.it/event/8401/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Automatic topography of complex data sets by accurate density estimation DTSTART;VALUE=DATE-TIME:20171011T090000Z DTEND;VALUE=DATE-TIME:20171011T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8411@ictp.it DESCRIPTION:\n Data sets can be considered an ensemble of realizations dra wn from a density distribution. Obtaining a synthetic description of this distribution allows to rationalize the underlying generating process and building human-readable models. In simple cases\, visualizing the distri bution in a suitable low-dimensional projection is enough to capture its main features but real-world data sets are often embedded in a high-dimens ional space.\n I present a procedure that allows to obtain such a synthet ic description in an automatic way with the only information of pairwise d ata distances (or similarities). This methodology is based on a reliable estimation of the intrinsic dimension of the dataset and the probability density function coupled with a modified Density Peaks clustering algorit hm.\n The final outcome of all this machinery working together is a hiera rchical tree that summarizes the main features of the data set and a clas sification of the data that maps which of these features they belong to.\n  \n\n//indico.ictp.it/event/8411/ LOCATION:ICTP Central Area\, 2nd floor\, SISSA Building\, via Beirut URL://indico.ictp.it/event/8411/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Learning through stochasticity in discrete neural net works DTSTART;VALUE=DATE-TIME:20171113T103000Z DTEND;VALUE=DATE-TIME:20171113T113000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8432@ictp.it DESCRIPTION:\n \n \n \n Stochasticity and limited precision of synap tic weights in neural network models are key aspects in both biological an d hardware modeling of learning processes.\n I'll introduce a neural ne twork model with stochastic binary weights that naturally gives prominence to exponentially rare dense regions of solutions.\n These solutions in dense regions have a number of desirable properties such as robustness an d good generalization performance\, while typical solutions are isolated a nd hard to find.\n Binary solutions of the standard perceptron problem are obtained from a simple gradient descent procedure on a set of real val ues parametrizing a probability distribution over the binary synapses.\n I'll present analytical and numerical results on the percepetron problem \, along with algorithmic extensions aimed at training discrete deep neura l networks.\n \n References:\n Baldassi\, Gerace\, Kappen\, Lucibe llo\, Saglietti\, Tartaglione\, and Zecchina.\n On the role of synaptic stochasticity in training low-precision neural networks. ArXiv:1710.09825 \n Baldassi\, Borgs\, Chayes\, Ingrosso\, Lucibello\, Saglietti\, and Z ecchina.\n Unreasonable effectiveness of learning neural networks: From accessible\n states and robust ensembles to basic algorithmic schemes. PNAS\, 2016.\n Baldassi\, Ingrosso\, Lucibello\, Saglietti\, and Zecch ina.\n Subdominant Dense Clusters Allow for Simple Learning and High Co mputational Performance in Neural Networks with Discrete Synapses. PRL\, 2 015.\n \n \n\n\n//indico.ictp.it/event/8432/ LOCATION:ICTP\, Central Area\, 2nd floor\, old SISSA building Via Beirut URL://indico.ictp.it/event/8432/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Statistical Physics-inspired models of biological net work: collective behavior in neuronal ensembles DTSTART;VALUE=DATE-TIME:20171123T103000Z DTEND;VALUE=DATE-TIME:20171123T113000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8444@ictp.it DESCRIPTION:\n In both cortices and sensory systems\, information is repre sented and transmitted through the correlated activity of large neuronal n etworks. Methods borrowed from Statistical Physics and Machine Learning ar e powerful tools for characterizing the collective behavior of large syste ms and thus offer promising approaches to understand the activity of neuro nal populations. In this talk I will show how the Maximum Entropy principl e\, applied to cortical in-vivo recording\, allows for comparing the popul ation behavior during wakefulness and deep sleep and eventually for identi fying cell-assemblies\, i.e. strongly co-activated groups of neurons that play a central role in memory consolidation. I will then use hidden layer models\, point processes and “experimental” linear response theory to account for the non-linear stimulus processing in sensory networks such as the retina. These approaches allow for constructing high performing model s of the retinal population response to visual stimuli and thus for charac terizing how a network of neurons can encode and transmit visual informati on.\n\n//indico.ictp.it/event/8444/ LOCATION:ICTP Central Area\, 2nd floor\, SISSA building URL://indico.ictp.it/event/8444/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Disentangling bacterial invasiveness from lethality i n an experimental host-pathogen system DTSTART;VALUE=DATE-TIME:20180108T103000Z DTEND;VALUE=DATE-TIME:20180108T113000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8470@ictp.it DESCRIPTION:\n Understanding virulence remains a central problem in human health\, pest control\, disease ecology and evolutionary biology. Bacteria l virulence is typically quantified by phenomenological indicators such as the LT50 (i.e. the time taken to kill 50% of an infected population). How ever\, virulence emerges as a result of complex processes that occur at di fferent stages: the pathogen needs to breach the primary host defenses\, f ind a suitable environment to replicate\, and finally express the virulenc e factors that cause lethality. It is well-known that pathogens exhibit a very broad spectrum of strategies to accomplish these three tasks\, yet\, phenomenological indicators such as the LT50 cannot distinguish the abilit y of the pathogen to invade the host from its ability to kill the host. He re\, we propose a physical host-pathogen theory that shows how to disentan gle colonization\, growth\, and pathogen lethality from the survival kinet ics of a host population. Experimental data from Caenorhabditis elegans ne matodes exposed to various human pathogens shows that host mortality becom es severe only once the pathogen population has reached its carrying capac ity within the host. In the talk\, I will present our theory and compare p redictions against experimental data.\n\n//indico.ictp.it/event/8470/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building\, Via Beirut URL://indico.ictp.it/event/8470/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Records of Entropy Production at the Nanoscale DTSTART;VALUE=DATE-TIME:20180109T130000Z DTEND;VALUE=DATE-TIME:20180109T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8473@ictp.it DESCRIPTION:\n The laws of thermodynamics can be extended to the nanoscale \, where fluxes are fluctuating quantities. Little is known about extreme- value statistics of thermodynamic fluxes characterising the most extreme d eviations from the average behaviours. Using Martingale theory\, we study statistics of the negative records of stochastic entropy production in non equilibrium steady states\, and derive universal inequalities for such dis tributions of records. Furthermore\, we explore the implications of our re sults in two non-equilibrium nanoscopic systems: single-electron transisto rs and molecular motors. We report on the experimental measurement of stoc hastic entropy production and of records of negative entropy in a metallic double dot under a constant external DC bias. Experimental results on the double dot confirm our theory and reveal a novel bound for the maximal he at that a mesoscopic system can absorb from its environment. We also explo re our results in active biological processes and find predictions for the maximal excursion of a molecular motor against the direction of an extern al force.\n\n//indico.ictp.it/event/8473/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building\, Via Beirut URL://indico.ictp.it/event/8473/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Fitness response relation of growing population: an application of large deviation theory and semi-Markov processes DTSTART;VALUE=DATE-TIME:20180111T100000Z DTEND;VALUE=DATE-TIME:20180111T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8481@ictp.it DESCRIPTION:\n Fitness is a macroscopic quantity that characterizes the po tential success of a growing population of organisms and cells.\n Fitness is also a major target or objective for controlling growth and expansion o f pathological bacteria and tumors by perturbing properties of individual cells such as replication rate\, death rate and so on.\n Better understand ing of the nature of growing populations and the controllability over them requires us to clarify the relation how the fitness is determined by thes e single-cell or single-organism level properties.\n To this end\, we inve stigate a multi-type age-structured population model in which the inter-di vision interval and phenotypic states of cells are accounted.\n By using a path integral formulation of the model and its contraction onto the pairw ise empirical quantities\, we derive a response relation of the fitness to the change in the several single-cell-level parameters.\n This work is an extension of our previous work with Dr. Yuki Sughiyama [1].\n  \n [1] Y. Sughiyama\, T.J. Kobayashi\, et al.\, Pathwise thermodynamic structure in population dynamics\, Phys. Rev. E. 2015.\n\n//indico.ictp.it/event/8481/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8481/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Cooperation vs. Competition in an Evolutionary Ecolog ical Framework DTSTART;VALUE=DATE-TIME:20180115T093000Z DTEND;VALUE=DATE-TIME:20180115T103000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8484@ictp.it DESCRIPTION:\n Abstract:\n In an ecological system pathogens often need to share their host with other pathogens\, and therefore compete for the res ources with different spreading strategies. They can interact in different manners: for a short period cooperation between pathogens can lead to fas ter and larger host occupation [1-7]. Spanish Flu and HIV are examples of such cases. The cooperation\, however\, can lead to death of the host popu lation and consequently also pathogens’ death. Therefore on a long run\, the cooperation strategy is not necessary the best. We propose and study an evolutionary game model in order to understand the co-evolutionary dyna mics of two co-infecting pathogens [8]. They have a common host and the ho st does not evolve on the same time scale as the pathogens. We consider tw o kind of disease species\, while each of them have two different strategi es: cooperation and defection. Agents (pathogens) accumulate a payoff\, ba sed on the history of their contagion records. This gives rise to two main scenarios: The first is when the disease infects an empty host. In this s cenario\, the pathogen does not meet any resistance\, and all the host res ources are available to him. But when the host is already occupied by anot her disease\, things become a bit more complicated. More specifically in t he second scenario\, there are four possible combinations of pairs of stra tegies: (C\, C)\, (C\, D)\, (D\, C) and (D\, D) corresponding to different payoffs\; A cooperator pathogen does not show any resistance to the conta gion by another disease and will share the host resources with it. However \, a defector entering a host populated by a cooperator will seize the maj ority of available resources. Considering Hawk-Dove game\, in a mean-field approximation\, we first show under which conditions cooperation may or m ay not be a meaningful strategy. Then we show how evolution affects spread ing dynamics\, and if and how any strategy can win. Finally we show how un derlying transmission and contact networks may promote both the spreading and the emergence of cooperation.\n Moreover\, we show non-trivial dynamic al effects of cooperation and competition in an ecological framework [9\, 10]\; we study two strains competing with each other for host resources in the presence of a third pathogen cooperating with both of them. We treat dynamics in a homogeneously mixed population by means of mean-field theory and stability analysis and also on complex transmission networks. We stud y the impact of cooperation on the outcome of the two-pathogen competition \, which can be quantified in terms of dominance of one competing pathogen or the co- circulation of both of them. We show that the presence of a th ird cooperating pathogen can alter the outcome of competition as it may fa vor the more cooperative pathogen over the more infectious one.\n [1] L. C hen\, F Ghanbarnejad\, W. Cai\, P. Grassberger\, EPL 104\, 5 (2013).\n [2] W. Cai\, L. Chen\, F. Ghanbarnejad\, P. Grassberger\, Nature Physics 11\, 936 (2015).\n [3] P. Grassberger\, L. Chen\, F. Ghanbarnejad\, W. Cai\, P hysical Review E 93\, 042316 (2016).\n [4] J. P. Rodríguez\, F. Ghanbarn ejad\, V. M. Eguíluz\, Frontiers in Physics\, V 5\, P 46 (2017).\n [5] S . Sajjadi\, F. Karimi\, M. R. Ejtehadi\, F. Zarei\, S. Moghimi-Araghi\, F. Ghanbarnejad\, “Coinfection in different time scales”\, manuscript in preparation. [6] L. Chen\, F. Ghanbarnejad\, D. Brockmann\, New J. Phys. 19\, 103041(2017).\n [7] J. P. Rodríguez\, F. Ghanbarnejad\, V. M. Egui ́luz\, “How mobility affect cooperative spreading diseases”\, manuscr ipt in preparation.\n [8] F. Ghanbarnejad\, K. Seegers\, A. Cardillo\, P. Hoevel\, “Evolutionary cooperation\, yes or no?”\, manuscript in prepa ration.\n [9] F. Pinotti\, F. Ghanbarnejad\, P. Hoevel\, C. Poletto\, “H ow to compete in presence of a cooperative agent”\, manuscript in prepar ation.\n [10] S. Meloni\, F. Ghanbarnejad\, Y. Moreno “coinfection on mu ltilayer networks”\, manuscript in preparation.\n\n//indico.ictp.it/even t/8484/ LOCATION:ICTP Central Area\, 2nd floor\, SISSA Building URL://indico.ictp.it/event/8484/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: A Finite-size Scaling Framework Uncovers the Covariat ions of Ecological Scaling Laws DTSTART;VALUE=DATE-TIME:20180216T100000Z DTEND;VALUE=DATE-TIME:20180216T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8521@ictp.it DESCRIPTION:\n Scaling laws in ecology are recurrent and pervasive pattern s observed in ecosystems\, intended both as functional relationships among ecologically-relevant quantities and the probability distributions that c haracterize their occurrence. Well-known examples include the Species-Area relationship (SAR)\, quantifying the increase of biodiversity with ecosys tem area\, and Kleiber’s law\, the allometric relation between organismi c size and metabolic rate. The interest in these laws lies in their intrin sic predictive power: how many species would go extinct if the ecosystem s hrinks to half its size? What is the mass of the largest organisms inhabit ing ecosystems of different extent? Are there more large-sized or small-si zed organisms and species? Scaling laws observed empirically often conform to power-laws\, A=Ba\, where a is the scaling exponent. Although their fu nctional form appears to be ubiquitous\, empirical scaling exponents may v ary with ecosystem type and resource supply rate.\n While ecological laws have been often studied independently\, simple heuristic reasonings show t hat they are linked. Such reasonings\, however\, do not allow accounting f or finite-size effects\, restricting the range for power-law behavior in f inite ecosystem due to ecological or biological constraints on organismic size\, or for other deviations from pure power-laws. These limits demand f or a different approach.  The ubiquity of power-laws and the presence of finite-size constraints suggest finite-size scaling theory as a useful too l in this context. A scaling hypothesis for the joint probability distribu tion of abundance and body mass of species inhabiting an ecosystem of fini te size is proposed and used to derive macroecological patterns. The hypot hesis is supported by a broad class of resource-limited community dynamics stochastic models. Precise linkages among ecological laws were derived fr om the proposed scaling hypothesis\, in the form of algebraic relationship s among scaling exponents. Such relationships rationalize the observed var iability of ecological exponents across ecosystems\, clarifying how change s in one ecological pattern affect the remaining ones. Predicted covariati ons were verified on empirical data. This model-free approach allows inves tigating the effects of different ecological or biological assumptions on the covariation of scaling exponents.\n\n//indico.ictp.it/event/8521/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8521/ END:VEVENT BEGIN:VEVENT SUMMARY:Joint MATH-QLS Seminar - A simple tool for complex problems: The a daptive interpolation method for the Wigner spiked model DTSTART;VALUE=DATE-TIME:20180227T130000Z DTEND;VALUE=DATE-TIME:20180227T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8551@ictp.it DESCRIPTION:\n There has been much progress recently in proving single-let ter formulas for the mutual information (or "free energy") in high-dimensi onal estimation and learning problems. Computing the mutual information is important in order to locate the various "phase transitions" occurring in such problems when the noise increases or the data becomes too scarce. It is also key in computing various optimal achievable errors. Unfortunately all existing methods are highly involved\, difficult to generalize and re stricted in their applicability. In this talk I'll present a new method\, called "adaptive interpolation method"\, that eliminates these barriers al l at once: It is much simpler\, very generic and able to tackle problems t hat were resisting until now. I will illustrate the method on a paradigmat ic model of high-dimensional estimation\, namely the "Wigner spiked model" (or "low-rank matrix factorization"). I will also briefly review some mod els that are now under full rigorous control thanks to this approach\, as well as very recent extensions to physics models such as the "ferromagneti c p-spin model on sparse random graphs"\, an open problem for decades for reasons that I'll mention.\n\n//indico.ictp.it/event/8551/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA Building\, via Beirut URL://indico.ictp.it/event/8551/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Entropy production in non-equilibrium systems DTSTART;VALUE=DATE-TIME:20180323T100000Z DTEND;VALUE=DATE-TIME:20180323T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8557@ictp.it DESCRIPTION:\n Systems out of equilibrium are characterized by a continuou s exchange of energy and matter with the environment through different mec hanisms\, thus producing entropy at the macroscopic level. We study the en tropy production of a discrete-state system with random transition rates a nd the injection of an external probability current. At stationarity\, the entropy production is shown to be composed by two contributions whose ex act distributions are evaluated in the large system size limit and close t o equilibrium. The first one is related to a generalized Joule's law\, whi le the second contribution has a Gaussian universal distribution which dep ends just on two topological parameters. In the second part of the talk we compare the entropy production of a Master Equation system with the one o f the corresponding continuum limit\, i.e. a Fokker-Planck equation. We de monstrate that the Seifert's formula for the entropy production provides j ust a lower bound for the exact entropy production. In fact\, this formula ought to be corrected keeping into account information about the microsco pic transition rates\, and this correction survives in the continuum limit . This effect of the coarse-graining has shown to be glaring in the simple example of a n-step random walk. The difference between discrete-state sy stems and continuous systems in evaluating the entropy production has been addressed also in comparing non-equilibrium steady states\, NESS\, and st ochastic pumping\, SP. We show that the entropy production of a system wit h SP is greater than the one produced by a system in a NESS\, generating t he same (time-averaged) probability current and exhibiting the same (time- averaged) probability distribution.\n\n//indico.ictp.it/event/8557/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8557/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Natural selection in compartmentalized environment wi th reshuffling DTSTART;VALUE=DATE-TIME:20180613T090000Z DTEND;VALUE=DATE-TIME:20180613T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8598@ictp.it DESCRIPTION:\n The emerging field of compartmentalized in vitro evolution\ , where selection is carried out by differential reproduction in each comp artment\, is a promising new approach to protein engineering. From a pract ical point of view\, it is important to know the effect of the increase in the average number of genotype bearing agents per compartment. This effec t is also interesting on its own in the context of primordial evolution in the hypothetical RNA world. The question is important as genotypes with d ifferent phenotypes in the same compartment share their fitness (the numbe r of produced copies) rendering the selection frequency-dependent. I will show the results of a theoretical investigation of this problem in the con text of selection dynamics for a simple model with an infinite population that is periodically redistributed among infinite number of identical comp artments\, inside which all molecules are copied without distinction with the success rate as a function of the total genomic composition in the com partment. Surprisingly\, with a linear selection function\, the selection process is slowed down only approximately inversely proportional to the av erage number of individuals per compartment. I will also demonstrate exact forms of the governing equations for some nonlinear selection functions. Finally\, I will expose an intriguing open problem of an apparent phase tr ansition for an exponential selection function seen in numerical experimen ts\, which is missed by the current infinite population theory.\n\n//indic o.ictp.it/event/8598/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building\, Via Beirut 2 URL://indico.ictp.it/event/8598/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Vectors of Chagas disease - determination of species by infrared spectroscopy and machine learning DTSTART;VALUE=DATE-TIME:20180614T090000Z DTEND;VALUE=DATE-TIME:20180614T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8607@ictp.it DESCRIPTION:\n Chagas disease\, caused by the parasite Trypanosoma cruzi\, is widespread in Latin America\, where the disease remains one of the maj or public health problems\, with an estimation of 8 million infected peopl e and 10 000 deaths per year. This condition is mostly transmitted by inse cts\, known as kissing bugs\, belonging to the Triatominae family (Hemipte ra)\, which are obligate haematophagous insects all their life. More than 150 species have been described in the world. While the majority lives in the wild\, some of them are highly associated to the human beings\, living in or around the dwellings. They typically stay hidden in the wall or roo f cracks of homes\, going out at night to feed on human blood. The correct determination of the species involved in the transmission of the disease is crucial to develop efficient control strategies. This can be achieved b y keys of determination for adult stages\, or by molecular technics. Both technics are time and/ or money consuming\, showing the needs of new ident ification methods\, especially for nymphal instars. These last years\, var ious publications demonstrated the potential of infrared spectroscopy in i nsect taxonomy. Bolivia is a highly endemic country for Chagas disease. Th e main vector\, Triatoma infestans\, lives on about 60% of the territory. In total\, 17 species of triatomines are reported in the country\, among t hem Triatoma sordida and Triatoma guasayana which are reported as secondar y vectors. These two species are sympatric and morphologically similar\, a nd so they are difficult to discriminate. The goal of this study was to de velop a classification model\, using living nymphal and adult stage of the se three species. 1293 spectra were taken in the invisible and near-infrar ed range. Different models were built\, using different pre-processing met hodologies of the spectra\, and different types of feature selection. The performance of each model was evaluated for each species. After their comp arison\, the best model was tested on a different set of specimens where i t showed a global accuracy of 97% (95-99%)\, an F1 score greater than 0.95 and a specificity greater than 0.94. This result shows that using infrare d spectroscopy is a good strategy to predict the triatomine species. It is the first investigation to report the ability to identify juvenile instar s\, and moreover with a single model together with the adult stage.\n\n//i ndico.ictp.it/event/8607/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building\, Via Beirut 2 URL://indico.ictp.it/event/8607/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Stability and organisation in macroscopic systems DTSTART;VALUE=DATE-TIME:20180629T090000Z DTEND;VALUE=DATE-TIME:20180629T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8615@ictp.it DESCRIPTION:\n Macroscopic systems are often endowed with a large number o f different scales and features. While controlling one single system often demands to properly describe all its complexity\, different systems can e xpress similar behaviours which are scale-invariant and depend on a minima l set of parameters.\n The last two features allow to reproduce these phen omena at the laboratory scale and describe them with a minimal analytical model\, thus characterising the observed phenomena in a very general way.\ n  \n Following this approach I focus on the stability of different macro scopic systems\, from the human scale (suitcases and bikes)\, to the geoph ysical (vortices and currents in the ocean) and astrophysical scales (accr etion disks).\n I show that in all these systems the macroscopic behaviour relies on\n the balance or inbalance between few parameters.\n Also\, dep ending on the relative importance of these parameters\, these systems can show the appearing of longevous and well organised patterns.\n  \n Finall y I introduce a new project which focuses on the complex structure of the arboreal nests built by Nasute termites. I aim to characterise these objec ts using the same minimal approach than before while facing the additional biological aspects. I will try to answer three connected questions: what are the biological functions of the macroscopic structure\, how it is buil t by many simply interacting agents\, and at which extent these two elemen ts are related to the form of the nest.\n  \n\n//indico.ictp.it/event/861 5/ LOCATION:ICTP Central Area\, old SISSA building URL://indico.ictp.it/event/8615/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Polymer under gradient fields DTSTART;VALUE=DATE-TIME:20180702T090000Z DTEND;VALUE=DATE-TIME:20180702T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8614@ictp.it DESCRIPTION:\n In this talk\, we first discuss the equilibrium properties of a linear polymer chain in a confined space e.g. the cone-shaped channel (entropy gradient) to understand the transport of biopolymers from pore e tc. In the second part\, we shall discuss the effects of the net gradient field arising due to the competition between the velocity gradient and sol vent quality gradient (e.g. chemical potential\, pH of the solvent\, conce ntration of ions etc) on the dynamics of polymers. Multi-factor gradients are essential components of many biological phenomena. Insight into the be havior of such systems is of fundamental importance in a wide spectrum of systems.\n  \n\n//indico.ictp.it/event/8614/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8614/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Hydrodynamic signatures of stationary Marangoni-drive n surfactant transport DTSTART;VALUE=DATE-TIME:20180705T090000Z DTEND;VALUE=DATE-TIME:20180705T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8611@ictp.it DESCRIPTION:Surfactant spreading at air-water interfaces is driven by flow setup by surface tension gradients (Marangoni stress) established by the surfactants themselves. We experimentally probe the nature of steady surfa ctant transport on the interface when the resulting flow is strongest in a thin boundary layer near the interface. In particular\, we present three experimental hydrodynamic signatures to distinguish between two limiting c ases\, viz. adsorption versus dissolution dominated transport\, without in voking the surfactant's physico-chemical properties. In a region much larg er than the surfactant source\, but much smaller than the interfacial area \, the steady-state fluid velocity assumes a self-similar form whose magni tude decays as a power-law with the distance from the source. We experimen tally demonstrate that this power-law possesses an exponent -3/5 in adsorp tion and -1 in dissolution dominated flow. Explicit measurement of boundar y layer and shear stress provide additional hydrodynamic signatures of sur factant transport mechanisms in the two limiting cases. We test this crite rion against two known surfactants\, Sodium Dodecyl Sulfate and Tergitol 1 5-S-9\, and apply the results to camphoric acid\, with unknown surface pro perties.\n\n//indico.ictp.it/event/8611/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA Building URL://indico.ictp.it/event/8611/ END:VEVENT BEGIN:VEVENT SUMMARY:Disease ecology: How to compete in a multi-pathogen system? DTSTART;VALUE=DATE-TIME:20180717T120000Z DTEND;VALUE=DATE-TIME:20180717T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8628@ictp.it DESCRIPTION:\n In an ecological system\, pathogens often need to share the ir host with other pathogens\, and therefore compete for the resources wit h different spreading strategies. Both cooperative and competitive interac tions in bacterial infections have been observed.  Here we first review r ecent theoretical studies addressed these two mechanisms separately and t hen study their combination and discuss about non-trivial dynamical effec ts can be expected to arise. Thus we study two strains competing with each other for host resources in the presence of a third pathogen cooperating with both of them. We first treat dynamics in a homogeneously mixed popula tion by means of mean-field theory and stability analysis. We study the im pact of cooperation on the outcome of the two-pathogen competition\, which can be quantified in terms of dominance of one competing pathogen or the co-circulation of both of them. We show that the presence of a third coope rating pathogen can alter the outcome of competition as it may favor the m ore cooperative pathogen over the more infectious one. We then consider mo re complex contact structures among hosts and perform computer simulations to study the evolution of the diseases.\n\n//indico.ictp.it/event/8628/ LOCATION:ICTP Central Area\, 2nd floor\,old SISSA building URL://indico.ictp.it/event/8628/ END:VEVENT BEGIN:VEVENT SUMMARY:Revealing modular architecture of the cortical networks DTSTART;VALUE=DATE-TIME:20180731T153000Z DTEND;VALUE=DATE-TIME:20180731T163000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8635@ictp.it DESCRIPTION:\n In natural systems time delays are inevitable due to the fi nite speed of the signal transmission over a distance\, and are not neglig ible if they are similar to the time scales of the observed dynamics. For example in the brain\, where the delays are on the same scale as the signa l operation\, they affect the brain performances and need to be considered . On the other hand\, many complex networks\, such as the cortical network s\, are hierarchical. It means that their communities may be further divid ed into sub-communities. Therefore\, because of the above-mentioned reason s\, the analysis of time-delayed dynamics on hierarchical networks may hel p to understand the interplay between brain structure and function. Based on extensive simulations in artificial and cortical networks with homogene ous time delays\, we uncovered that for a fixed coupling strength\, by cha nging time delay different regions of coherent\, multistable and incoheren t dynamics are revealed. We show that in a hierarchical network\, by trans ition from incoherent to coherent states\, different topological scales of the network are revealed respectively in a non transient behavior. In add ition\, we find that by considering bimodal distribution of the time delay s the parameter regions corresponding to the multistabe dynamics will be e xtended. Finally we show that the results are connected to the global slow dynamics of the brain.\n\n//indico.ictp.it/event/8635/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8635/ END:VEVENT BEGIN:VEVENT SUMMARY:Trap Models and their Connection to more Realistic Glasses DTSTART;VALUE=DATE-TIME:20180905T090000Z DTEND;VALUE=DATE-TIME:20180905T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8677@ictp.it DESCRIPTION:\n \n At low temperatures\, the dynamic s of glasses suffer a dramatic slowing down. The system becomes stuck in m etastable configurations called traps\, which are rarely abandoned\, throu gh a process called activation. The Trap Model (TM)\, that de- scribe the motion between different traps\, provides a simplified framework to un- de rstand activated dynamics. Even though signs of TM-like behavior were foun d in realistic systems\, it is not clear (i) whether the dynamics of most models is the one predicted by the TM\, (ii) to what extent the TM descrip tion applies to other glasses and (iii) which are the relevant features fo r this dynamics to be found. We show that the TM description does apply to other glassy models\, such as the Ran- dom Energy Model [1]\, if one defi nes the traps dynamically\, through the time series of the energy [2]. We then extend our analysis to systems with correlated energy levels\, and se e that the trap behavior holds as long as the correlations are weak [3]. O nce the correlations are strong enough\, as it happens in a Hamiltonian sy stem\, it is unclear whether the dynamic behavior of the glass can be comp letely described through the simple picture provided by the TM.Comparing D ynamics of Glasses with Deep Neural Networks \n\n Time permitting\, a seco nd section on Deep Learning is presented. We study dynam- ics and energy ( or loss) landscape of feedforward Deep Neural Networks [4]\, with an empha sis on their Mean Square Displacement\, and find that they exhibit aging o n a finite time scale. Later\, they start diffusing at the bottom of the l andscape. Further\, we argue that the slow dynamics is due rather to flat directions in the landscape than to barrier crossing\, and we show evidenc e of an under- to over-parametrized phase transition. \n\n References \n\n 1.     [1]  M. Baity-Jesi\, G. Biroli & C. Cammarota\, J. Stat. Mech . (2018) 013301. \n\n 2.     [2]  C. Cammarota & E. Marinari\, Phys. Rev. E 92 010301(R) (2015). \n\n 3.     [3]  M. Baity-Jesi\, A. Achar d-de Lustrac & G. Biroli\, Phys. Rev. E 98\, 012133 (2018). \n\n 4.      [4]  M. Baity-Jesi\, L. Sagun\, M. Geiger\, S. Spigler\, G. Ben-Arous\ , C. Cammarota\, Y. LeCun\, M. Wyart & G. Biroli\, PMLR 80:324-333\, 2018 (ICML 2018). \n\n  \n\n \n \n\n  \n\n \n \n\n//in dico.ictp.it/event/8677/ LOCATION:ICTP Central Area\, Second floor\, old SISSA building URL://indico.ictp.it/event/8677/ END:VEVENT BEGIN:VEVENT SUMMARY:Towards physics of biological action and perception DTSTART;VALUE=DATE-TIME:20180917T090000Z DTEND;VALUE=DATE-TIME:20180917T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8670@ictp.it DESCRIPTION:\n Living systems differ from inanimate objects in their abili ty to unite basic laws of nature into chains and clusters leading to new s table and pervasive relations among physical variables and involving new p arameters. They produce actions by modifying these parameters. Some of suc h “biological laws of nature” have been described including those unde rlying the control of voluntary movements. This approach allows exploring stability of action and perception\, which looks next to miraculous given the continuously changing intrinsic states of the body and the changing en vironment. Stability of action and perception is viewed as the formation o f a stable low-dimensional manifold\, uncontrolled manifold for action and iso-perceptual manifold for perception\, in high-dimensional spaces of el ements. Over the past years\, we have developed a method of uncontrolled m anifold-based analysis of action stability in spaces of elemental performa nce variables and control variables for the effectors. These studies have shown\, in particular\, that humans are able to modify stability of steady states in preparation to action and that control of action stability is g rossly impaired in certain neurological disorders. I will also review a co uple of recent experimental studies testing some of the predictions of the scheme for stable perception. One of them explored force illusions induce d by muscle vibration. The other study explored perception of elemental va riables during a multi-element action.\n\n//indico.ictp.it/event/8670/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8670/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Establishment of large-scale patterns by short-range interactions in biological tissues DTSTART;VALUE=DATE-TIME:20181025T090000Z DTEND;VALUE=DATE-TIME:20181025T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8774@ictp.it DESCRIPTION:Biological tissues represent fascinating dynamics leading to t he formation of structures and patterns at larger scales compared to cell size. It is a big challenge to understand how cells communicate with each other to coordinate their dynamics and create orders. As an example\, cell s sort out into distinct populations\, keeping a sharp straight interface between them. On the other side\, cells can express polarities and the pol arity field reveals long distance patterns. In this talk\, I would like to present physical models to address some mechanisms that contribute to est ablishment of stable patterns. First\, we will discuss how local interacti ons of cells at the interface influence its shape and mechanics. In the se cond case\, I present a theoretical model to describe dynamics of polarity fields. Our model generalizes classical spin models by incorporating effe ctive transport of polarity molecules between neighboring cells. We study how the intercellular transport qualitatively affects dynamics and leads t o the formation of non-trivial patterns.\n\n//indico.ictp.it/event/8774/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8774/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS GUEST SEMINAR: Inference of latent growth-related states of ce lls from cellular lineage trees DTSTART;VALUE=DATE-TIME:20181113T173000Z DTEND;VALUE=DATE-TIME:20181113T183000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8799@ictp.it DESCRIPTION:\n Individual cells are phenotypically heterogeneous even with the same genetic background. The heterogeneity endows the cells with diff erent growth capabilities. This difference in turn works as the driving fo rce of the natural selection among the cells\, and the population graduall y adapt to the environment as the population level.\n Therefore\, quantifi cation of the growth capabilities from data is crucial for understanding h ow the evolutionary learning of the cells is shaped and regulated in a pop ulation.\n To this end\, we propose an Expectation-Maximization (EM) based algorithm to infer the growth capability as hidden states of cells from c ellular lineage trees. We modified the EM algorithm to account for the tre e structure of the lineage tree and resolved the sampling bias problem (su rvivorship bias) that inevitably appears in an evolving population.\n By a pplying our method to a lineage tree data of E.coli\, we identified hidden states of the cells that may be linked to the slowing changing activity o f the cells. This method may contribute to facilitating our quantitative u nderstanding on the Darwinian adaptation operating at the cellular level.\ n\n//indico.ictp.it/event/8799/ LOCATION:ICTP Adriatico Guest House - Kastler Lecture Hall URL://indico.ictp.it/event/8799/ END:VEVENT BEGIN:VEVENT SUMMARY:Joint CMSP/QLS Seminar: The Quantum Boltzmann Machine DTSTART;VALUE=DATE-TIME:20181115T130000Z DTEND;VALUE=DATE-TIME:20181115T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8773@ictp.it DESCRIPTION:\n We propose to generalise classical maximum likelihood learn ing to density matrices. As the objective function\, we propose a quantum likelihood that is related to the cross entropy between density matrices. We apply this learning criterion to the quantum Boltzmann machine (QBM)\, previously proposed by Amin et al.. We demonstrate for the first time lear ning a quantum Hamiltonian from quantum statistics using this approach. Fo r the anti-ferromagnetic Heisenberg and XYZ model we recover the true grou nd state wave function and Hamiltonian. The second contribution is to appl y quantum learning to learn from classical data. Quantum learning uses in addition to the classical statistics also quantum statistics for learning. These statistics may violate the Bell inequality\, as in the quantum case . Maximizing the quantum likelihood yields results that are significantly more accurate than the classical maximum likelihood approach in several ca ses. We give an example how the QBM can learn a strongly non-linear proble m such as the parity problem. The solution shows entanglement\, quantified by the entanglement entropy.https://arxiv.org/abs/1803.11278\n  \n  \n\ n//indico.ictp.it/event/8773/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8773/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Complex Systems: From networks to microbes DTSTART;VALUE=DATE-TIME:20181129T100000Z DTEND;VALUE=DATE-TIME:20181129T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8805@ictp.it DESCRIPTION:\n In this seminar\, consisting of two parts\, we present an o utline of our recent work in the area of networks and microbial systems.\n We propose a new edge based network metric and show its importance for al l networks. We also show how networks can be successfully applied to a wid e variety of problems ranging from information retrieval and non-invasive diagnostics to structural optogenetics and adaptive landscapes.\n Eliminat ing mycobacterial diseases through prevailing antibiotic-based strategies has had limited success\, due to the rise of drug-resistant strains. Phage therapy is an interesting potential alternative. We show how a strategic coupling of experimental data -- with theoretical techniques drawn from de lay differential equations\, graph theory\, discrete math and statistics - - sheds valuable insight into secondary host lethality and phage resistanc e in such systems.\n Lastly\, we provide a brief glimpse of other ongoing experimental and theoretical activity in our lab.\n\n//indico.ictp.it/even t/8805/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8805/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Wild Speculation Seminar - "What makes life sciences differen t from physics?" DTSTART;VALUE=DATE-TIME:20181219T150000Z DTEND;VALUE=DATE-TIME:20181219T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8826@ictp.it DESCRIPTION:\n \n Why is it that we can describe the physical properties of macroscopic objects of zillions of degrees of freedom with just a handf ul of variables? Why isn't it possible to describe\, in general\, living s ystems (a cell\, an organism\, an ecology or an economy) in the same way?  \n \n This blackboard talk claims to answer these questions (specially the second one) from a fundamental perspective based on statistical mechan ics and information theory. The format encourages the most critical attitu de from participants from all backgrounds.\n\n\n  \n\n  \n\n//indico.ict p.it/event/8826/ LOCATION:ICTP Central Area\, 2nd floor\, Old SISSA building URL://indico.ictp.it/event/8826/ END:VEVENT BEGIN:VEVENT SUMMARY:A few problems in physics of hearing DTSTART;VALUE=DATE-TIME:20190215T100000Z DTEND;VALUE=DATE-TIME:20190215T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8886@ictp.it DESCRIPTION:\n Hair cells are the receptor cells that enable the transduct ion of mechanical stimuli into the nervous system of hearing and balance o rgans of vertebrates. The name of these cells owes to the specialized orga nelle located at their tip—the hair bundle. The organelle's complex stru cture and operation\, which is enhanced by an active process\, are crucial for an organ's sensory function. In my talk I will tell about a few open physical problems related to the development and operation of the hair bun dle.\n\n//indico.ictp.it/event/8886/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8886/ END:VEVENT BEGIN:VEVENT SUMMARY:Mathematical modelling of red blood cell development DTSTART;VALUE=DATE-TIME:20190222T100000Z DTEND;VALUE=DATE-TIME:20190222T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8888@ictp.it DESCRIPTION:\n A requisite to understand the general principles underlying transcriptional regulation is to define its most important players (i.e. transcription factors and cofactors) quantitatively in absolute terms. Suc h global and quantitative knowledge is necessary to define the range of po ssibilities (i.e. the general context) within which TFs perform their func tions from interacting to forming complexes\, to being recruited to specif ic genes in specific nuclear environment. Furthermore\, competition betwee n TFs (for binding sites) is thought to be an important mechanism for driv ing cellular differentiation with changes in the stoichiometry between LS- TFs being considered as the central mechanism underlying cell fate choice. \n  \n We measure RNA-seq (transcriptomic) expression and SRM (proteomic) of a time series hematopietic stem cells differentiation into red blood c ells. To study the data\, first we perform a bioinformatics analysis of ou r samples to identify the most relevant genes. Second\, we build a Gene Re gulatory Network (GRN) of those genes based on previous knowledge and corr elations in our time series. Third\, we transform that GRN into a system o f Ordinary Differential Equations to obtain a quantitative model. Fourth\, we perform some knock-down experiments to validate our model. Our results show new interactions between genes and an accurate quantitative descript ion of red blood cell development.\n\n//indico.ictp.it/event/8888/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8888/ END:VEVENT BEGIN:VEVENT SUMMARY:CANCELLED - The thermodynamic uncertainty relation DTSTART;VALUE=DATE-TIME:20190226T160000Z DTEND;VALUE=DATE-TIME:20190226T170000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8894@ictp.it DESCRIPTION://indico.ictp.it/event/8894/ LOCATION:ICTP Adriatico Guest House - Kastler Lecture Hall URL://indico.ictp.it/event/8894/ END:VEVENT BEGIN:VEVENT SUMMARY:Algorithmic synthesis of complex biomolecules DTSTART;VALUE=DATE-TIME:20190319T133000Z DTEND;VALUE=DATE-TIME:20190319T143000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8907@ictp.it DESCRIPTION:\n \n \n \n An algorithm converts inputs to correspondin g unique outputs through a sequence of actions. Algorithms are used as met aphors for complex biological processes such as organismal development. He re we make this metaphor rigorous for the synthesis of the branched carboh ydrates known as glycans. These molecules play a key role in conveying cel lular identity and self/non-self information across all domains of life\; for example\, the famous A/B blood group antigens are glycans. In eukaryot ic cells (including human cells) glycans are synthesized by collections of enzymes in a factory-line system of compartments known as the Golgi appar atus. The Golgi can stochastically convert a single input molecule to a he terogeneous set of possible outputs\; yet in living cells the observed div ersity of the outputs is very small. Here we resolve this paradox by borro wing from the theory of algorithmic self-assembly. For a large class of st ochastic models\, given an input and a target output we either prove that the output cannot be algorithmically synthesized from the input\, or expli citly construct a succession of Golgi compartments that achieves this synt hesis. Our theoretical analysis allows us to infer the causes of heterogen eity in real glycan datasets.\n \n \n\n\n//indico.ictp.it/event/8907/ LOCATION:ICTP Adriatico Guest House - Kastler Lecture Hall URL://indico.ictp.it/event/8907/ END:VEVENT BEGIN:VEVENT SUMMARY:Bioimpedance and bone fracture detection DTSTART;VALUE=DATE-TIME:20190410T120000Z DTEND;VALUE=DATE-TIME:20190410T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8916@ictp.it DESCRIPTION:\n All tissues of the human body have electrical properties. T hese are applied to understand some physiological processes and anatomical structures. We show our research about the detection of bone fractures by bioimpedance measurement. We deal with the development of a non-invasive technology to this object: model and simulations\, electronic development and clinical measurements.\n\n  \n\n//indico.ictp.it/event/8916/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8916/ END:VEVENT BEGIN:VEVENT SUMMARY:Cooperative AI: Modelling and Resolving Complex Social Dilemmas us ing Deep Reinforcement Learning DTSTART;VALUE=DATE-TIME:20190416T080000Z DTEND;VALUE=DATE-TIME:20190416T090000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8928@ictp.it DESCRIPTION:\n Social dilemmas are ubiquitous in the world of human social interactions. The human ability to solve complex spatially and temporally extended social dilemmas is a key hallmark of our intelligence. Conversel y\, the cumulative nature of social tasks may provide an automatic path to wards increasingly powerful cognitive abilities\, under appropriate constr aints. In this talk\, I will argue for the ubiquity of social dilemmas fro m some simple principles\, and show how complex spatially and temporally e xtended dilemmas may be solved by including appropriate inductive biases i n state-of-the-art deep reinforcement learning agents. This approach simul taneously advances the fields of behavioural economics and artificial inte lligence. I shall end with an outlook on how these fields might interact m ore closely in the future\, both to improve our understanding of multi-age nt dynamics in human society\, and to construct generally intelligent arti ficial systems.\n\n//indico.ictp.it/event/8928/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/8928/ END:VEVENT BEGIN:VEVENT SUMMARY:Microbial motility and bacterial gene network dynamics DTSTART;VALUE=DATE-TIME:20190429T090000Z DTEND;VALUE=DATE-TIME:20190429T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8937@ictp.it DESCRIPTION:\n In this talk I will present a panoramic view of the results of some of the most recent projects within my group. This includes: study of motility of the bacterium E. coli in a porous medium with different de grees of porosity\; characterization of the motility patterns of the paras ite responsible for Chagas' disease\; and study of the dynamic behavior of the tryptophanase gene regulatory network in Escherichia coli. I also pla n to discuss possible future interdisciplinary research projects arising f rom the presented results. In this last regards\, I'm very much interested in having the audience feedback.\n\n//indico.ictp.it/event/8937/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8937/ END:VEVENT BEGIN:VEVENT SUMMARY:Active flow of low-density pedestrian crowds DTSTART;VALUE=DATE-TIME:20190507T123000Z DTEND;VALUE=DATE-TIME:20190507T133000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8942@ictp.it DESCRIPTION:\n The dynamics of pedestrian crowds share deep connections wi th the statistical physics of active matter.\n Pedestrians move following own will and objectives: this amounts in huge variability in the motion\, observable even in diluted conditions. Despite such individual unpredictab ility\, ensemble-level universal physical features emerge and encompass co mmon and rare fluctuations of both the solo and the interacting dynamics. Reaching a quantitative understanding of these features is a major scienti fic challenge with deep societal impact\, e.g. in the design of civil infr astructures or of crowd management measures.\n  \n We investigate from ob servational experiments held in real-life settings (stations\, festivals\, museums) and in diluted conditions statistical features of pedestrian mot ion. We leverage on datasets including millions of trajectories acquired w ith and without external influencing stimuli (i.e.\, crowd control measure s\, like signage or visual cues)\, via home-made high-fidelity tracking sy stems. On this basis\, we quantify the PDFs of individual velocity\, posit ion\, body rotation and mutual-contact-avoidance "social" forces - possibl y in dependence on stimuli. We propose an active-Brownian particle model o f the dynamics based on Langevin-like equations statistically quantitative .\n  \n  \n This work is in collaboration with: J. Meeusen\, C. Lee\, A. Muntean\, R. Benzi\, F. Toschi\n  \n\n//indico.ictp.it/event/8942/ LOCATION:Common Area ICTP\, Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8942/ END:VEVENT BEGIN:VEVENT SUMMARY:Universal Biology in Adaptation and Evolution: Multilevel Consiste ncy\, Dimension Reduction\, and Fluctuation-Response Relationship DTSTART;VALUE=DATE-TIME:20190513T120000Z DTEND;VALUE=DATE-TIME:20190513T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8945@ictp.it DESCRIPTION:\n A macroscopic theory for cellular states with steady-growth is presented\, based on consistency between cellular growth and molecular replication\, as well as robustness of phenotypes against perturbations. Adaptive changes in high-dimensional phenotypes are shown to be restricted within a low-dimensional slow manifold\, from which a macroscopic law for cellular-states is derived\, as is confirmed by adaptation experiments of bacteria under stress. Next\, the theory is extended to phenotypic evolut ion\, leading to proportionality between phenotypic responses against gene tic evolution and by environmental adaptation.  Evolutionary relevance of slow modes in controlling high-dimensional phenotypes is discussed. Last\ , if I have time\, transition from exponential-growth to stationary phases is investigated as the breakdown of steady-growth.  A general law betwee n the starvation time and lag-time to recover the cellular growth is discu ssed.\n  \nReferences\n \n\n  Kaneko K.\, Life: An Introduction to Com plex Systems Biology\, Springer (2006)\n \n  K. Kaneko\, C.Furusawa\, T. Yomo\, "Macroscopic phenomenology for cells in steady-growth state"\, Phy s.Rev.X(2015) 011014\n \n  C. Furusawa\, K. Kaneko "Global Relationships in Fluctuation and Response in Adaptive Evolution"\, J of Royal Society I nterface 12(2015)\, 20150482.\n \n  C. Furusawa\, K. Kaneko " Formation of Dominant Mode by Evolution in Biological Systems” Phys. Rev. E 97(201 8)042410\n \n  K. Kaneko\, C. Furusawa “Macroscopic Theory for Evolvin g Biological Systems Akin to Thermodynamics”\, Annual Rev. Biophys. (201 8) 47\, 273-290\n \n  Y. Himeoka\, K. Kaneko (2017). Theory for transiti ons between exponential and stationary phases: universal laws for lag time . Physical Review X\,(2017) 7\, 021049\n\n\n//indico.ictp.it/event/8945/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8945/ END:VEVENT BEGIN:VEVENT SUMMARY:Monomer-dimer models on random graphs DTSTART;VALUE=DATE-TIME:20190514T120000Z DTEND;VALUE=DATE-TIME:20190514T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8949@ictp.it DESCRIPTION:\n I will introduce monomer-dimer models\, an example of Stati stical Mechanics models on graphs characterized by hard-core interaction. A great contribution in this field was given by Heilmann and Lieb\, who pr oved the absence of phase transitions under very general hypothesis.\n  \ n\n In the second part of the talk I will focus on monomer-dimer models on sparse random graphs\, which can be solved exactly by means of a cavity m ethod and rigorously thanks to a family of correlation inequalites.\n\n//i ndico.ictp.it/event/8949/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8949/ END:VEVENT BEGIN:VEVENT SUMMARY:Workshop on Martingales in Finance and Physics DTSTART;VALUE=DATE-TIME:20190524T070000Z DTEND;VALUE=DATE-TIME:20190524T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8947@ictp.it DESCRIPTION:\n Martingales are paradigmatic stochastic processes used in p robability theory and finance as models of fair financial markets with no net gain or loss. The tool of martingales has been shown to be crucial in financial analysis with fruitful applications in the context of option pri cing. In physics\, martingales have not been explored much yet. Over the l ast few years\, a new martingale formulation of the non-equilibrium thermo dynamics of small systems has been developed within the field of stochasti c thermodynamics. This workshop will establish an interdisciplinary dialog ue on the use of martingales as “skeleton key” to the fluctuating worl d of finance and stochastic thermodynamics\, and explore novel connections between these fields.\n  \n\n Invited Speakers:\n Giorgia Callegaro - Pa dua U.\n Shamik Gupta - Ramakrishna Mission Vivekananda U.\n Gonzalo Manza no - Scuola Normale\, Pisa & ICTP\n Izaak Neri - King's College\, London\n Fulvio Ortu -  U. Bocconi\, Milan\n Édgar   Roldán -  ICTP\n Feder ico Severino -  USI\, Lugano\n\n \n  \n\n//indico.ictp.it/event/8947/ LOCATION:ICTP Room D\, SISSA building\, Level -1 URL://indico.ictp.it/event/8947/ END:VEVENT BEGIN:VEVENT SUMMARY:Higher-order memory and inertia effects in Onsager-Machlup theory of stochastic fluctuations DTSTART;VALUE=DATE-TIME:20190604T120000Z DTEND;VALUE=DATE-TIME:20190604T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8964@ictp.it DESCRIPTION:\n In two seminal papers [1\, 2] Onsager and Machlup introduce d the celebrated theory of stochastic fluctuations for thermodynamic syste ms.\n Whereas their first contribution [1] proposes a variational principl e that leads to a simple Langevin dynamics of first order in time\, the se cond [2] extends this formalism to more complex systems that include inert ia and memory effects via a second-order stochastic equation. I will prese nt a higher-order generalization of the Onsager-Machlup theory and discuss its applications to equilibrium and nonequilibrium dynamics of currents.\ n  [1] L. Onsager and S. Machlup\, Phys. Rev. 91\, pp. 1505-12\, 1953.\n [2] S. Machlup and L. Onsager\, Phys. Rev. 91\, pp. 1512-15\, 1953.\n\n//i ndico.ictp.it/event/8964/ LOCATION:ICTP Common Area\, 2nd floor\, SISSA building URL://indico.ictp.it/event/8964/ END:VEVENT BEGIN:VEVENT SUMMARY:Nanotechnology-based advanced sensors for surveillance\, stealth a nd healthcare DTSTART;VALUE=DATE-TIME:20190610T120000Z DTEND;VALUE=DATE-TIME:20190610T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8955@ictp.it DESCRIPTION:\n Sensors are an integral part of any instrumentation\, mecha nical assembly\, automobile engineering\, heavy engineering\, drug-deliver y vehicles\, national surveillance gadgets or any electromagnetic applicat ion unit\, such as antennas and communication electronics. A need for smar t\, adaptive\, miniaturized\, extremely sensitive\, selective and rapid se nsor is always on anvil. Nanomaterials are offering promising options to i mprove the conventional sensors and to develop advanced sensors which woul d be adaptive\, faster\, selective and more precise. Nanotechnology has be en dominating applied research frontiers for more than a decade now. The r esearch has an emphasis on exploring novel materials with exotic propertie s\, which are attributed to their nano-size-regimes. Typically explored ex amples are metals\, oxides and chalcogenides\, in their pure and composite forms.\n Through this presentation a brief outline on the progress of sci ence and technology\, in the domain of sensors will be given. Frequency de tections\; low-field and frequency (electric and magnetic fields and ultra -low-frequency signals)\, high frequency (radio frequencies in Radar range ) detections will be discussed. Biosensors being developed for disease dia gnostics and smart drug delivery vehicles would be also elaborated in this presentation. Various approaches used in the investigators laboratory wou ld be elaborated\; namely\, the optical fiber approach\, metamaterial appr oach and conventional resistive approach. The relationships of the obtaine d properties would be associated with the physics and chemistry at nano-le vel and their energy dynamics for sensing a particular physical parameter. \n\n//indico.ictp.it/event/8955/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8955/ END:VEVENT BEGIN:VEVENT SUMMARY:Nanotechnology-based advanced sensors for surveillance\, stealth and healthcare DTSTART;VALUE=DATE-TIME:20190612T120000Z DTEND;VALUE=DATE-TIME:20190612T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8978@ictp.it DESCRIPTION:\n Sensors are an integral part of any instrumentation\, mecha nical assembly\, automobile engineering\, heavy engineering\, drug-deliver y vehicles\, national surveillance gadgets or any electromagnetic applicat ion unit\, such as antennas and communication electronics. A need for smar t\, adaptive\, miniaturized\, extremely sensitive\, selective and rapid se nsor is always on anvil. Nanomaterials are offering promising options to i mprove the conventional sensors and to develop advanced sensors which woul d be adaptive\, faster\, selective and more precise. Nanotechnology has be en dominating applied research frontiers for more than a decade now. The r esearch has an emphasis on exploring novel materials with exotic propertie s\, which are attributed to their nano-size-regimes. Typically explored ex amples are metals\, oxides and chalcogenides\, in their pure and composite forms.\n Through this presentation a brief outline on the progress of sci ence and technology\, in the domain of sensors will be given. Frequency de tections\; low-field and frequency (electric and magnetic fields and ultra -low-frequency signals)\, high frequency (radio frequencies in Radar range ) detections will be discussed. Biosensors being developed for disease dia gnostics and smart drug delivery vehicles would be also elaborated in this presentation. Various approaches used in the investigators laboratory wou ld be elaborated\; namely\, the optical fiber approach\, metamaterial appr oach and conventional resistive approach. The relationships of the obtaine d properties would be associated with the physics and chemistry at nano-le vel and their energy dynamics for sensing a particular physical parameter. \n\n//indico.ictp.it/event/8978/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8978/ END:VEVENT BEGIN:VEVENT SUMMARY:Emergent Hyperbolic Network Geometry and Dynamics DTSTART;VALUE=DATE-TIME:20190614T120000Z DTEND;VALUE=DATE-TIME:20190614T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8960@ictp.it DESCRIPTION:\n \n \n Simplicial complexes naturally describe discrete t opological spaces. When their links are assigned a length they describe di screte geometries. As such simplicial complexes have been widely used in q uantum gravity approaches that involve a discretization of spacetime. Rece ntly they are becoming increasingly popular to describe complex interactin g systems such a brain networks or social networks. In this talk we presen t non-equilibrium statistical mechanics approaches to model large simplici al complexes. We propose the simplicial complex model of Network Geometry with Flavor (NGF)\, we explore the hyperbolic nature of its emergent geome try and their relation with quantum statistics. Finally we reveal the rich interplay between Network Geometry with Flavor and dynamics. We investiga te the percolation properties of NGF using the renormalization group findi ng BKT and discontinuous phase transitions depending on the dimensionality simplex. We also comment on the synchronization properties of NGF and the emergence of frustrated synchronization.\n \n\n\n  \n\n//indico.ictp.it/ event/8960/ LOCATION:ICTP Meeting room 201\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/8960/ END:VEVENT BEGIN:VEVENT SUMMARY:Efficiency of finite-time heat engines: from linear-irreversible t o low-dissipation behavior DTSTART;VALUE=DATE-TIME:20190619T133000Z DTEND;VALUE=DATE-TIME:20190619T143000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8981@ictp.it DESCRIPTION:\n There is a great interest in understanding the optimal feat ures  of the performance of finite-time thermal machines. Efficiency at m aximum power and the coefficient of performance at maximum cooling power\, are some of the useful criteria that are studied.\n Universal features of these quantities are observed in many different models\, such as those ba sed on endoreversibility assumption\, low-dissipation assumption\, stochas tic heat engines\, and even quasi-static models. We discuss a model of hea t engine within the framework of linear irreversible thermodynamics.\n The model runs in a finite cycle time. We show that the model can be mapped t o a low-dissipation model whereby the entropy generated in a heat-exchange step\, is inversely proportional to the duration of the process. The opti mal features of the efficiency will be discussed.\n\n//indico.ictp.it/even t/8981/ LOCATION:ICTP Central Area\, 2nd floor\, SISSA building URL://indico.ictp.it/event/8981/ END:VEVENT BEGIN:VEVENT SUMMARY:Mechanical response of cells to substrate topography DTSTART;VALUE=DATE-TIME:20190627T090000Z DTEND;VALUE=DATE-TIME:20190627T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8992@ictp.it DESCRIPTION:\n Cells sense and respond to topographical\, chemical\, and m echanical characteristics of their substrate. Engineered substrates are in creasingly being developed that exploit these physical attributes to direc t cell responses and therefore control cell behavior toward desired applic ations.\n There are very few methods available for robust and accurate mod eling that can predict cell deformations and force propagations along the cell media. Here\, we developed a unifying computational framework to crea te a multi-component cell model to predict cell behavior. The model which we call it “virtual cell model” has the capability to predict changes in whole cell and its nucleus characteristics [1]. The changes in shape an d elongation of the membranes\, force distribution\, and chromatin contact map are predicted on a range of substrates.\n Modeling data are correlate d with cell culture experimental outcomes in order to confirm the applicab ility of the model and demonstrating the ability to reflect the qualitativ e behavior of mesenchymal stem cells [1\,2].\n Here we show controlled cel l migration is possible just by proper design of the substrate.\nReference s:[1] ACS Nano\, 2017\, 11 (9)\, pp 9084–9092\, DOI: 10.1021/acsnano.7b0 3732\n[2] Adv. Funct. Mater. 2018\, 1707378\, DOI: 10.1002/adfm.201707378\ n \n\n//indico.ictp.it/event/8992/ LOCATION:ICTP Central Area\, 2nd floor\, SISSA building URL://indico.ictp.it/event/8992/ END:VEVENT BEGIN:VEVENT SUMMARY:Construction\, estimation\, diagnostics\, and extensions of the Ha wkes-type point-process models in analysing social and natural science dat a DTSTART;VALUE=DATE-TIME:20190710T093000Z DTEND;VALUE=DATE-TIME:20190710T093000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8986@ictp.it DESCRIPTION:\n Point process modeling has been widely used as a powerful s tatistical tool in describing\, investigating\, and forecasting social\, e conomic\, and natural phenomena. Especially\, the Hawkes-type models descr ibe the interaction between discrete events: each event\, no matter whethe r it is a background event or an excited event\, influences the occurrence of the future events\, through triggering (“encouraging” of) events\, according to some probability rules. Since the Hawkes models can be used to investigate the interaction between discrete events\, they became more and more popular in many fields\, such as crime data analysis\, economics\ , and seismology.\n In this talk\, using the ETAS model for earthquake occ urrences and a nonparametric Hawkes model for crime data as the illustrati ve examples\, I will explain the methods of stochastic declustering and st ochastic reconstruction. Stochastic declustering helps us to realize the f amily-tree structure among the events in a stochastic manner. Stochastic r econstruction is a technique for model diagnostics and later is generalize d to a nonparametric method for estimating the branching characteristics f or more general branch process models\, by incorporating other smoothing t echniques. The stochastic reconstruction method can be validated through t he viewpoints of martingale theory. I will show its powers by the examples of extending two existing Hawkes-type models in earthquakes and crimes da ta analysis.\n  \n\n//indico.ictp.it/event/8986/ LOCATION:ICTP Central Area\, 2nd floor\, SISSA building URL://indico.ictp.it/event/8986/ END:VEVENT BEGIN:VEVENT SUMMARY:Stochastic bifurcation in a turbulent swirling flow DTSTART;VALUE=DATE-TIME:20190712T093000Z DTEND;VALUE=DATE-TIME:20190712T093000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8987@ictp.it DESCRIPTION:\n We report the experimental evidence of the existence of a r andom strange attractor in a fully developed turbulent swirling ow. By def ining a global observable which tracks the asymmetry in the of angular mom entum imparted to the ow\, we can reconstruct the associated turbulent att ractor and then follow its route towards chaos. We further show that the e xperimental attractor can be modeled by stochastic Duffing equations\, tha t match the quantitative properties of the experimental flow\, namely\, th e number of quasi-stationary states and transition rates among them\, the effective dimensions\, and the continuity of the Lyapunov exponents. Such properties can be recovered neither using deterministic models nor using s tochastic differential equations based on effective potentials obtained by inverting the probability distributions of the experimental global observ ables. A random map extracted from the experimental time series of the tur bulent ow exhibits qualitatively same stochastic bifurcation structure as the experimentally observed transitions. Our findings open the way to low- dimensional dynamical system modeling of systems featuring a large number of degrees of freedom and multiple quasi-stationary states.\n  \n\n//indi co.ictp.it/event/8987/ LOCATION:ICTP Central Area\, 2nd floor\, URL://indico.ictp.it/event/8987/ END:VEVENT BEGIN:VEVENT SUMMARY:A Formal Framework for Intuitive Probabilistic Reasoning DTSTART;VALUE=DATE-TIME:20190715T093000Z DTEND;VALUE=DATE-TIME:20190715T093000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8988@ictp.it DESCRIPTION:\n I propose a formal framework for intuitive probabilistic re asoning (IPR)\, which aims to capture the informal\, yet logical\, process by which individuals justify beliefs about uncertain claims\, as in legal arguments\, mathematical conjecture\, scientific hypothesizing\, and comm on sense reasoning. The proposed model of IPR is independent of\, but can be viewed as complementary to\, Bayesian and related personalist decision theories. In this talk\, I highlight the key ideas of this new system incl uding its philosophical grounding in mathematical intuitionism (Brouwer\, 1981\; Heyting\, 1971\; Dummett\, 2000) and its formalization in intuition istic Martin- Löf type theory (Martin-Löf\, 1984\, 1987\, 1996) and ho motopy type theory (Univalent Foundations Program\, 2013). In addition to describing the formal system (syntax and semantics) and its immediate cons equences\, I compare and contrast with more conventional theories of decis ion under uncertainty\, in particular Bayesian decision theory\, Dempster –Shafer theory\, and imprecise probability.\n  \n\n//indico.ictp.it/eve nt/8988/ LOCATION:ICTP Central Area\, 2nd floor\, SISSA building URL://indico.ictp.it/event/8988/ END:VEVENT BEGIN:VEVENT SUMMARY: Kinetic Uncertainty Relation DTSTART;VALUE=DATE-TIME:20190715T130000Z DTEND;VALUE=DATE-TIME:20190715T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9001@ictp.it DESCRIPTION:\n We start from a recent theorem of microscopic thermodynamic s\, pictorially called thermodynamic uncertainty relation (TUR)\, which is part of the more general framework of thermodynamic inequalities. More sp ecifically\, the above-mentioned theorem states that the ratio between the average squared of an out-of-equilibrium current and its variance is boun d from above by the entropy production. However\, it was shown that this r elation can be obtained in a very general way using the Ito-Dechant formal ism and the Kullback-Leibler divergence that\, for some particular cases\, can be linked to entropy production. We will hence use the latter to obta in a new non-equilibrium inequality for systems modelled by continuous tim e Markov chains\, i.e. performing a simple perturbation on the system's dy namic (namely a linear rescaling of the transition rates) we calculate the Kullback-Leibler divergence that arises  from this procedure and show th at it is proportional to the frenesy (or dynamical activity) of the system .\n This final result has been called Kinetic Uncertainty Relation (KUR). Finally\, we investigate the differences  between the TUR and the KUR and analyse the regimes in which they give tighter constraints.\n\n//indico.i ctp.it/event/9001/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9001/ END:VEVENT BEGIN:VEVENT SUMMARY:Microscopic chaos\, fractals and diffusion: From simple models tow ards experiments DTSTART;VALUE=DATE-TIME:20190718T093000Z DTEND;VALUE=DATE-TIME:20190718T103000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8989@ictp.it DESCRIPTION:\n A century ago Einstein developed a theory of diffusion that is based on the assumption of stochasticity for a Brownian particle. On a molecular level\, however\, diffusion is generated by Newton’s determin istic equations of motion\, which typically are highly nonlinear. In my ta lk I will review\, in a very tutorial manner\, a theoretical approach that derives diffusion from first principles by starting from microscopic dete rministic chaos. Simple models ranging from a deterministic random walk on the line to chaotic diffusion in potentials on periodic lattices will be studied analytically and by computer simulations. Surprisingly\, the diffu sion coefficients of these models turn out to be highly irregular\, fracta l functions of physical control parameters while simple random walk theory predicts monotonicity. This is due to memory in the deterministic single- particle dynamics. I will argue that this phenomenon should be seen in exp eriments on small systems\, such as molecular diffusion in zeolite nanopor es and diffusion of electrons in artificial graphene.\n  \n\n//indico.ict p.it/event/8989/ LOCATION:ICTP Central Area\, 2nd floor\, SISSA building URL://indico.ictp.it/event/8989/ END:VEVENT BEGIN:VEVENT SUMMARY:Thermodynamic notions for time-delayed stochastic processes DTSTART;VALUE=DATE-TIME:20190718T123000Z DTEND;VALUE=DATE-TIME:20190718T133000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9010@ictp.it DESCRIPTION:\n Stochastic thermodynamics provides a consistent description of a wide class of Langevin systems\, but the Markov assumption is often essential. While some non-Markovian systems have been studied in great det ail\, the important case of feedback control with a discrete delay is stil l insufficiently understood [1\,2]. The thermodynamical description of non linear delayed systems turns out to be particularly challenging.\n In this talk\, I will consider the paradigmatic example of a Brownian particle in a doublewell potential subject to time-delayed feedback in its non-equili brium steady state. We use different strategies to tackle the technical ch allenges arising due to the delay\, such as closure schemes for the infini te Fokker-Planck hierarchy [3\,4] and a Markovian embedding technique [4\, 5].\n We find an unavoidable heat flow induced by the control (even in the absence of any additional driving)\, i.e.\, the feedback inevitably cools down or heats up the system [1]. Interestingly\, the heat flow is enhance d in the limit of small delay times\, which is a phenomenon related to ent ropy pumping. We further compute the total entropy production of the super system\, consisting of the particle plus the memory of the controller. By varying the information capacity of the memory device\, we can consider d ifferent Gamma-distributed delays\, including the limit of an infinitely s harp (delta-peaked) kernel\, i. e.\, discrete delay. We identify an entrop ic contribution due to the memory\, which is a consequence of the unidirec tional information flow induced by the feedback-controller. This contribut ion diverges in the case of error-free measurement\, as well as in the cas e of a perfectly sharp (delta-peaked) time delay.\n  \n [1] Loos & Klapp\ , Sci. Rep. 9\, 2491 (2019)\n [2] Rosinberg\, Munakata\, Tarjus\, PRE 91\, 042114 (2015) [3] Loos & Klapp\, PRE 96\, 012106 (2017)\n [4] Loos & Klap p\, ArXiv:1903.02322 (2019)\n [5] Puglisi & Villamaina\, EPL 88\, 30004 (2 009)\n\n//indico.ictp.it/event/9010/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9010/ END:VEVENT BEGIN:VEVENT SUMMARY:Can we predict weather and project climate using machine learning techniques? DTSTART;VALUE=DATE-TIME:20190723T093000Z DTEND;VALUE=DATE-TIME:20190723T093000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8990@ictp.it DESCRIPTION:\n Recurrent neural networks have been recently introduced to predict the behavior of chaotic systems. Surprisingly\, they provide relia ble forecasts up to the predictability limit without knowing the underlyin g equations of the systems. Here we focus on the predictability of rare tr ajectories leading to extreme events. We focus on different chaotic system s\, ranging from toy models of the atmospheric dynamics\, up to the re-for ecast of sea-level pressure and temperature fields issued by ERA Interim a nd NCEP reanalysis datasets. Our results show that: i) the machine learnin g techniques can make reliable weather forecasts of global atmospheric fie lds up to 48-72h with a time step of 3-6h ii) long time forecasts do not d iverge\, they produce reliable climatology of the sea-level pressure and t emperature fields iii) we can evaluate the coherence of extreme events gen erated via the machine learning with respect to the observed ones. \n\n \n \n\n//i ndico.ictp.it/event/8990/ LOCATION:ICTP Central Area\, 2nd floor\, SISSA building URL://indico.ictp.it/event/8990/ END:VEVENT BEGIN:VEVENT SUMMARY:American options with stochastic interest rates DTSTART;VALUE=DATE-TIME:20190723T120000Z DTEND;VALUE=DATE-TIME:20190723T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9013@ictp.it DESCRIPTION://indico.ictp.it/event/9013/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9013/ END:VEVENT BEGIN:VEVENT SUMMARY:The ergodicity problem in economics DTSTART;VALUE=DATE-TIME:20190726T093000Z DTEND;VALUE=DATE-TIME:20190726T093000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-8991@ictp.it DESCRIPTION:\n The ergodicity problem queries the equality or inequality o f time averages and expectation values. I will trace its curious history\, beginning with the origins of formal probability theory in the context of gambling and economic problems in the 17th century. This is long before e rgodicity was a word or a known concept\, which led to an implicit assumpt ion of ergodicity in the foundations of economic theory. 200 years later\, when randomness entered physics\, the ergodicity question was made explic it. Over the past decade my collaborators and I have asked what happens to foundational problems in economic theory if we export what is known about the ergodicity problem in physics and mathematics back to economics. Many problems can be resolved. Following an overview of our theoretical and co nceptual progress\, I will report on a recent experiment that strongly sup ports our view that human economic behavior is better described as optimiz ing time-average growth rates of wealth than as optimizing expectation val ues of wealth or utility of wealth. \n\n \n \n\n//indico.ictp.it/event/8991/ LOCATION:ICTP Central Area\, 2nd floor\, SISSA building URL://indico.ictp.it/event/8991/ END:VEVENT BEGIN:VEVENT SUMMARY:Microscopic Langevin model for the modified Fürth equation DTSTART;VALUE=DATE-TIME:20190731T123000Z DTEND;VALUE=DATE-TIME:20190731T133000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9012@ictp.it DESCRIPTION:\n Recent experimental research on the mobility of living cell s indicate that their diffusion is not completely described by Fürth's eq uation since they display a normal diffusive regime before the onset of th e ballistic motion. Moreover\, results from a cellular Potts model simulat ion for the motion of cells in two dimensions present the same type of beh avior and a modification of Fürth's equation has been proposed to fit the mean squared displacements (MSD) found both experimentally and computatio nally. Here\, we propose a microscopic Langevin model to explain this anom aly in the MSD: we show that a polarization direction for the persistent m otion as well as at least two time scales are needed. The model reproduces well the MSD obtained from the modified Fürth equation and is robust to changes in the microscopic dynamics of the polarization as a function of c ell migration speed\, which is in agreement with experimental data which s uggests that persistence of trajectories is coupled to cell migration spee d mediated by actin flows.\n\n//indico.ictp.it/event/9012/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9012/ END:VEVENT BEGIN:VEVENT SUMMARY:Controlling active spinners using vortex lattices DTSTART;VALUE=DATE-TIME:20190909T090000Z DTEND;VALUE=DATE-TIME:20190909T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9054@ictp.it DESCRIPTION:\n Actively spinning particles attracted much attention in rec ent years\, mostly from a point of view of their ability to self-assemble on the liquid surface. On the other hand\, the control of surface flows us ing surface waves has led to interesting analogies with optical traps for microparticles and ultracold atoms. In this talk I will present recent exp eriments on the confinement of active spinners within a wave-driven vortex lattice\, also referred to as liquid metamaterial. I will discuss propert ies of the surface flow generated by two orthogonal standing waves. The sp inners can be efficiently manipulated in such a flow to either occupy stat ionary orbits within a metamaterial cell\, or they can be moved between th e cells. I will discuss forces acting on the spinners in such flows and wi ll show how multiple spinners interact to form stable self-organised confi gurations.\n\n//indico.ictp.it/event/9054/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9054/ END:VEVENT BEGIN:VEVENT SUMMARY:Control of bacteria attachment and biofilm growth using surface w aves DTSTART;VALUE=DATE-TIME:20190909T100000Z DTEND;VALUE=DATE-TIME:20190909T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9055@ictp.it DESCRIPTION:\n Surface adherent microbial consortia\, biofilm growth\, is a predominant growth mode for bacteria and other microorganisms in nature. While there is considerable literature related to the negative aspects of the biofilms including chronic infections and industrial biofouling\, the y can also be harnessed as biotechnological tools such as in waste water t reatment and remediation of environmental pollutions.\n  \n We are intere sted in identifying fluid conditions favouring or discouraging bacteria at tachment and formation of biofilms and evaluating the role of the surface waves and vibration in the biofilm formation.\n  \n Our results show that the formation of the biofilms is strongly affected by the surface wave. D eterministic flows in a thin layer of nutrient medium promote the growth o f patterned biofilms while chaotic and turbulent fluid motion reduces it. The settlement of bacteria and the locations of biofilms differ from the s edimentation patterns of inactive bacteria and of the passive microparticl es\, suggesting that the environmental factors dominate over hydrodynamic advantages. Wave-driven flows control the delivery routes of nutrients\, o xygen and of the bacteria communication agents thus allowing to shape the biofilm development and to control macroscopic biofilm patterns. The resul ts offer efficient tools to encourage or discourage the attachment of bact eria and allow shaping the biofilms using low-frequency waves.\n\n//indico .ictp.it/event/9055/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9055/ END:VEVENT BEGIN:VEVENT SUMMARY:Chaotic diffusion for a family of area preserving mappings DTSTART;VALUE=DATE-TIME:20190910T090000Z DTEND;VALUE=DATE-TIME:20190910T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9056@ictp.it DESCRIPTION:\n The chaotic diffusion for a family of Hamiltonian mappings whose angles diverge in the limit of vanishingly action is investigated by using the solution of the diffusion equation. The system is described by a two-dimensional mapping for the variables action\, $I$\, and angle\, $\\ theta$ and controlled by two control parameters: (i) $\\epsilon$\, control ling the nonlinearity of the system\, particularly a transition from integ rable for $\\epsilon=0$ to non-integrable for $\\epsilon\\ne 0$ and\; (ii) $\\gamma$ denoting the power of the action in the equation defining the a ngle. For $\\epsilon\\ne 0$ the phase space is mixed and\n chaos is presen t in the system leading to a finite diffusion in the action characterized by the solution of the diffusion equation. The analytical solution is then compared to the numerical simulations showing a remarkable agreement betw een the two procedures. For the chaotic dynamics far apart from the period ic islands\, normal diffusion is observed. The scenario changes significan tly when the dynamics passes near stability regions where anomalous diffus ion dominates over the dynamics\, stickiness is present and a temporary br eak of ergodicity is observed.\n\n//indico.ictp.it/event/9056/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9056/ END:VEVENT BEGIN:VEVENT SUMMARY:Memristive networks DTSTART;VALUE=DATE-TIME:20190923T120000Z DTEND;VALUE=DATE-TIME:20190923T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9063@ictp.it DESCRIPTION:Nanoscale components often exhibit memory\, both quantum and c lassical. Resistors are no exception (already at the classical level) and the use of memristors for a variety of purposes is currently under scrutin y. Applications range from machine learning on chip to compact and passive memory devices using crossbar arrays. In this talk we discuss the nitty-g ritty mathematical aspects of memristive circuits in the analog regime (no CMOS)\, their connection to the Ising model\, and plausible future direct ions beyond crossbars.\n \n\n//indico.ictp.it/event/9063/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9063/ END:VEVENT BEGIN:VEVENT SUMMARY:Understanding the underlying mechanisms of pattern formation and c ellular aggregation DTSTART;VALUE=DATE-TIME:20190924T090000Z DTEND;VALUE=DATE-TIME:20190924T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9081@ictp.it DESCRIPTION:\n Cellular self-assembly and organization are fundamental ste ps for the development of biological tissues. Within the framework of a ce llular automata model\, we address how an ordered tissue pattern spontaneo usly emerges from a randomly migrating single cell population without the influence of any external cues. The simulations also capture several dynam ical properties of growing aggregates\, such as\, the rate of cell aggrega tion and non-monotonic growth of the aggregate area which show a good agre ement with the existing experimental observations. We further investigate the ruggedness of the growing structures by evaluating the fractal dimensi on to get insights into the complexity of tumorous tissue growth which wer e hitherto unexplored. Efficient process of collective migration of cells can be dictated by incorporating different factors like mechanical\, chemi cal and physical contact interaction present in the environment of cells f or tissue development. Our work exploits the mathematical features of diff erent kind of interacting forces and focuses on the contribution of these factors to the large scale organization.\n\n//indico.ictp.it/event/9081/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9081/ END:VEVENT BEGIN:VEVENT SUMMARY:Reading neural responses to avoid information loss DTSTART;VALUE=DATE-TIME:20190926T090000Z DTEND;VALUE=DATE-TIME:20190926T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9137@ictp.it DESCRIPTION:\n Classic studies show that in many species – from leech an d cricket to primate – responses of neural populations can be quite succ essfully read out using a measure of neural population activity termed the population vector. However\, despite its successes\, detailed analyses ha ve shown that the standard population vector discards substantial amounts of information contained in the responses of a neural population\, and so is unlikely to accurately describe how signal communication between parts of the nervous system. I will describe recent theoretical results showing how to modify the population vector expression in order to read out neural responses without information loss\, ideally. These results make it possi ble to quantify the contribution of weakly tuned neurons to perception. I will also discuss numerical methods that can be used to minimize informati on loss when reading out responses of large neural populations.\n  \n\n// indico.ictp.it/event/9137/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9137/ END:VEVENT BEGIN:VEVENT SUMMARY:Active matter far from equilibrium: Work\, dissipation and phase t ransitions DTSTART;VALUE=DATE-TIME:20191009T090000Z DTEND;VALUE=DATE-TIME:20191009T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9138@ictp.it DESCRIPTION:\n Active matter is a paradigm of soft materials made of a lar ge number of interacting agents\, ranging from colonies of bacteria to ass emblies of biomimetic micro-swimmers\, where individual self-propulsion le ads to dynamics and structures without any equilibrium equivalent. The dis sipation at the basis of activity offers the opportunity to design smart m aterials\, for instance to extract work with unprecedented protocols\, for which generic guiding principles are still lacking. Moreover\, controllin g dissipation allows one to select which order emerges at large scale by p romoting dynamical phase transitions\, whose properties are yet to be stud ied.\n First\, I will present design strategies for active engines which e xploit specifically nonequilibrium effects\, such as the autonomous motion of asymmetric obstacles and the lack of an equation of state in active fl uids. I will discuss how to optimize their efficiency in terms of the prot ocol details and\, when possible\, compare their performances with thermal engines. Second\, I will examine the emergence of spontaneous order when changing dissipation. Using large deviation techniques\, I will describe u nexpected phase transitions\, for instance towards a collective motion des pite the lack of aligning interactions\, and rationalize the microscopic m echanisms triggering such collective effects.\n\n//indico.ictp.it/event/91 38/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9138/ END:VEVENT BEGIN:VEVENT SUMMARY:Migration Italian Style\, 1977-2018 DTSTART;VALUE=DATE-TIME:20191023T090000Z DTEND;VALUE=DATE-TIME:20191023T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9170@ictp.it DESCRIPTION:Today\, in Spain\, Portugal\, Italy\, Malta and Greece\, the i ncidence of foreigners on the population is entirely comparable to that of the more traditional European immigration countries. Only forty years ago \, however\, in these five Southern European countries the size of the for eign presence was decidedly modest. Inverting a secular trend\, the migrat ion balance with foreign countries has become positive since the 1970s. Bu t after the migration boom of the beginning of the 21st century\, a sudden and sharp decline was observed in the following years of economic crisis. \nThis talk has a dual purpose. First it describes forty years of Italian migrations (1977-2018)\, systematically distinguishing Center-North from S outhern Italy and connecting it with the migratory history of previous dec ades. We show how the "stop and go" of migrations can be interpreted in th e light of the pull factors determined by structural changes in demography and in the labor market. Secondly\, it identifies the persistent and stru ctural peculiarities that have shaped the foreign presence in Italy\, buil ding a model very different from that of Central and Northern Europe.\n\n/ /indico.ictp.it/event/9170/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9170/ END:VEVENT BEGIN:VEVENT SUMMARY:Heat transfer at the nanoscale: A classical molecular dynamics app roach DTSTART;VALUE=DATE-TIME:20191029T100000Z DTEND;VALUE=DATE-TIME:20191029T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9192@ictp.it DESCRIPTION:\n  \n\n Today\, thermal transport management in nanoelectron ic devices is an essential factor in their design\, efficiency\, and lifet ime. Moreover\, thermal conductivity plays a crucial role in the figure of merit of the thermoelectric devices. Because of fundamental aspects and e ssential technological applications\, nanoscale heat transport has emerged as an interdisciplinary area of research involving condensed matter physi cs\, material sciences\, electrical and mechanical engineering\, etc. Heat transfer at the nanoscale is substantially different from that at the mac roscale and is governed by the mean free path of heat carriers in material s. In this presentation\, I will talk about the unusual thermal behavior o f materials at the nanoscale utilizing the classical molecular dynamics si mulation approach. I'll also show how mechanical strain\, doping\, functio nalization\, and grain size could influence the thermal conductivity of na nostructures. Moreover\, the effects of boundaries and interfaces on the t hermal transport of nanostructures will be discussed.\n  \n\n//indico.ict p.it/event/9192/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9192/ END:VEVENT BEGIN:VEVENT SUMMARY:Dynamic Brain-state allocation in health and neurodegeneration DTSTART;VALUE=DATE-TIME:20191031T130000Z DTEND;VALUE=DATE-TIME:20191031T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9198@ictp.it DESCRIPTION:\n At the moment\, when a patient presents with clinical sympt oms of Parkinson’s or Alzheimer's disease there is very little the clini cal community can do other than  manage the symptoms of what are inevitab le degenerative conditions.  Alterations in the neural state (or states) that eventually will produce functional and behavioural consequences had b een set in motion perhaps 20 years earlier\, and provoked the search for p roper biomarkers in asymptomatic individuals. The reorganisation of functi onal brain networks\, either due to slow changes in the relative contribut ion/wiring of brain areas (plasticity) or due to fast modulation of their causal interactions (effective connectivity)\, mask the early stages of ne urodegeneration\, until alterations cannot be compensated.  Until recentl y\, approaches to brain function in health and disease mostly did not take into account the fact that the brain is highly dynamic\, and that studyin g brain function as a series of static states may not be the most effectiv e way to encounter early biomarkers for neurodegeneration. Rather\, the cl ues might lay in the path or trajectory by which functional brain networks evolve and transition over time or their dwell time irregularities. Metho dologically\, estimating the evolution of the non-stationary and non-linea r brain is challenging. Most solutions apply a fixed predetermined moving window\, over which functional connectivity is estimated\, thus missing th e explicit modelling of the dynamical regimes\, across multiple temporal l ength-scales\, over which the brain networks are assumed to switch\, from few multi-seconds to seconds. The assumption here is that early stages of neurodegeneration could be detected and differentiated from normal ageing as alterations in the dynamical repertoires that the brain networks explor e\, either at rest or during task performance. These alterations would be reflected either in terms of the probability of the brain switching betwee n network configurations\, the duration it spends in particular configurat ions\, or in terms of the sequence or path of configurations.\n In this ta lk I will discuss advantages and disadvantages of current tools and presen t a conceptual framework for modelling brain dynamics that bridges between bottom-up data driven models and top-down biophysical models.\n  \n\n//i ndico.ictp.it/event/9198/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9198/ END:VEVENT BEGIN:VEVENT SUMMARY:Microtubule pivoting driven by cross-linking proteins leads to spi ndle assembly DTSTART;VALUE=DATE-TIME:20191127T100000Z DTEND;VALUE=DATE-TIME:20191127T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9199@ictp.it DESCRIPTION:\n During mitosis\, microtubules form a spindle\, which is res ponsible for the segregation of chromosomes. In yeast cells\, the spindle has a rod-like structure and is made of microtubules emanating from two po les connected by cross-linking proteins. Microtubules self-organize into p arallel or antiparallel bundles\, depending on whether they grow from the same or two different poles and our goal here is understanding how such st ructures form. Our model includes thermally driven angular pivoting of mic rotubules around the poles and elastic forces between them mediated by cro ss-linking proteins\, which can detach to and detach from microtubules\, a s well as move along them. The solutions of our model imply that the rando m motion of the microtubules allows them to find their pair\, while the sh ort-range interactions caused by the cross-linking proteins align them int o bundles. Parallel bundling can occur in the presence of either passive c rosslinkers or plus-end directed motors\, while the formation of antiparal lel bundles requires minus-end directed motors. The model predicts the ave rage bundling time\, which is in agreement with our experimental measureme nts. Additionally\, for the case of antiparallel bundle formation\, the mo del predicts that the velocity of the microtubules gliding along each othe r is the same as the velocity at which the motors move along the microtubu les\, and this was also confirmed experimentally. In conclusion\, the main contributors to the formation of microtubule bundles are angular diffusio n of microtubules around the poles allowing them to come into contact and short-range forces caused by cross-linking proteins that align them.\n\n// indico.ictp.it/event/9199/ LOCATION:ICTP Central area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9199/ END:VEVENT BEGIN:VEVENT SUMMARY:Generalized entropies in information theory and statistical mechan ics: A brief review DTSTART;VALUE=DATE-TIME:20191217T100000Z DTEND;VALUE=DATE-TIME:20191217T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9256@ictp.it DESCRIPTION:\n Over the past decades there has been a plausible interest f or the generalized entropies that go beyond Boltzmann-Gibbs one. Thus\, th e information theory community has been primarily devoted to the analysis of their usefulness in the contexts of source and channel coding\, as well as on the axiomatic characterization which approves their usefulness as t he measures of information. On the other hand\, the statistical physics co mmunity focused its efforts into developing the generalized thermostatisti cs and its applications in non-Boltzmann-Gibbs distributed systems and in the systems which exhibit non-extensive behavior.\n  \n In this talk we w ill review some of the existing definitions of generalized entropies\, tak ing into account their information-theoretic and thermostatistical propert ies. We will consider their applications in communication theory\, combina torial optimization problems and complex systems and neural networks model ing. Recent results will be reviewed and open questions and possible resea rch directions will be presented.\n\n//indico.ictp.it/event/9256/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9256/ END:VEVENT BEGIN:VEVENT SUMMARY:Interplay between competitive and cooperative interactions in a th ree-player pathogen system DTSTART;VALUE=DATE-TIME:20200114T133000Z DTEND;VALUE=DATE-TIME:20200114T143000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9273@ictp.it DESCRIPTION:\n In ecological systems\, heterogeneous interactions between pathogens take place simultaneously. This occurs\, for instance\, when two pathogens cooperate\, while at the same time multiple strains of these pa thogens co-circulate and compete. Notable examples include the cooperation of HIV with antibiotic-resistant and susceptible strains of tuberculosis\ , or some respiratory infections with Streptococcus pneumoniae strains.\n Models focusing on competition or cooperation separately fail to describe how these concurrent interactions shape\n the epidemiology of such disease s. We studied this problem considering two cooperating pathogens\, where o ne pathogen is further structured in two strains. The spreading follows a susceptible-infected-susceptible process and the strains differ in transmi ssibility and extent of cooperation with the other pathogen. We combined a mean-field\n stability analysis with stochastic simulations on networks c onsidering both well-mixed and structured populations. We observed the eme rgence of a complex phase diagram\, where the conditions for the less tran smissible\, but more cooperative strain to dominate are non-trivial\, e.g. non-monotonic boundaries and bistability. Coupled with community structur e\, the presence of the cooperative pathogen enables the co-existence betw een strains by breaking the spatial symmetry and dynamically creating diff erent ecological niches.\n These results shed light on ecological mechanis ms that may impact the epidemiology of diseases of public health concern.\ n\n//indico.ictp.it/event/9273/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9273/ END:VEVENT BEGIN:VEVENT SUMMARY:Towards Deep Network of Memory DTSTART;VALUE=DATE-TIME:20200213T100000Z DTEND;VALUE=DATE-TIME:20200213T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9317@ictp.it DESCRIPTION:\n Memory is one of the core brain functions of animals. Memor ies are distributed across cortical structures and have various time scale s\, ranging e.g. from sensory memory\, short-term memory to long term memo ry. How  memories are organized as a deep network is a critical question both for neuroscience and artificial intelligence research.\n\n  \n\n In this talk\, I will review recent progress in understanding the spatial mem ory circuits in the brain\, and discuss neural network models of cognitive map. By understanding the dynamics and computational mechanisms of the me mory circuits\, it is possible to find a path to understand and create int elligence.\n\n \n \n\n//indico.ictp.it/event/9317/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9317/ END:VEVENT BEGIN:VEVENT SUMMARY:Modelling strategies to organize healthcare workforce during pande mics: application to COVID-19 DTSTART;VALUE=DATE-TIME:20200403T120000Z DTEND;VALUE=DATE-TIME:20200403T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9340@ictp.it DESCRIPTION:\n Protection of healthcare workforce is of paramount relevanc e for the care of infected and non-infected patients in the setting of a p andemic such as coronavirus disease 2019 (COVID-19). Healthcare workers ar e at increased risk to become infected because of contact to infected pati ents\, infected co-workers and their community outside the hospital. \n\n The ideal organisational strategy to protect the healthcare workforce in a situation in which social distancing cannot be maintained at the workplac e remains to be determined. In this study\, we have mathematically modelle d strategies for the employment of hospital workforce with the goal to sim ulate health and productivity of the workers. Therefore\, deterministic mo dels were extended to account for stochastic influences potentially occurr ing in rather small populations. \n\n The models were also designed to det ermine if desynchronization of medical teams by dichotomizing the workers may protect the workforce. Our studies model workforce productivity depend ing on the infection rate\, the presence of reinfection and the efficiency of home office. As an application example\, we apply our theory to the ca se of coronavirus disease 2019 (COVID-19). The results of the models revea l that a desynchronization strategy in which two medical teams work altern ating for 7 days reduces the infection rate of the healthcare workforce. I n case of immunity to the infectious agent this affect is mainly relevant at early stages of the pandemic. This effect is independent on infection r ates and increases the overall workforce productivity under certain situat ions.\n\n  \n\n Link: \n\n https://www.medrxiv.org/content/10.1101/2020.0 3.23.20041863v1.article-metrics\n\n  \n\n\n\n//indico.ictp.it/event/9340/ LOCATION:Zoom Meeting Meeting ID: 475-819-702 URL://indico.ictp.it/event/9340/ END:VEVENT BEGIN:VEVENT SUMMARY:Asymptotic errors for convex penalized linear regression beyond Ga ussian matrices and extension to the generalized linear model DTSTART;VALUE=DATE-TIME:20200414T120000Z DTEND;VALUE=DATE-TIME:20200414T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9345@ictp.it DESCRIPTION:\n \n \n \n We consider the problem of learning a coeffi cient vector x0 ∈ RN from noisy linear observations y=Fx0+w∈RM  in th e high dimensional limit M\,N → ∞ with α ≡ M/N fixed. We provide a rigorous derivation of an explicit formula —first conjectured using heur istic methods from statistical physics— for the asymptotic mean squared error obtained by penalized convex regression estimators such as the LASSO or the elastic net\, for a class of very generic random ma- trices corres ponding to rotationally invariant data matrices with arbitrary spectrum. T he proof is based on a convergence analysis of an oracle version of vector approximate message-passing (oracle-VAMP) and on the properties of its st ate evolution equations. Our method leverages on and highlights the link b etween vector approximate message-passing\, Douglas-Rachford splitting and proximal descent algorithms\, extending previous results obtained with i. i.d. matrices for a large class of problems. We illustrate our results on some concrete examples and show that even though they are asymptotic\, our predictions agree remarkably well with numerics even for very moderate si zes. \n \n We then show how the same proof can be extended to the gen eralized linear model using an oracle version of generalized vector approx imate message passing (oracle-GVAMP) and a more elaborate convergence proo f based on Lyapunov arguments from control theory. \n \n \n\n\n  \n\n//i ndico.ictp.it/event/9345/ LOCATION:ZOOM MEETING Meeting ID: 475-819-702 URL://indico.ictp.it/event/9345/ END:VEVENT BEGIN:VEVENT SUMMARY:What really is a financial instrument? Insights from cross-impact DTSTART;VALUE=DATE-TIME:20200417T120000Z DTEND;VALUE=DATE-TIME:20200417T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9350@ictp.it DESCRIPTION:\n Cross-impact\, namely the fact that on average buy (sell) t rades on a financial instrument induce positive (negative) price changes i n other correlated assets\, can be measured from abundant\, although noisy \, market data. In this paper we propose a principled approach that allows to perform model selection for cross-impact models\, showing that symmetr ies and consistency requirements are particularly effective in reducing th e universe of possible models to a much smaller set of viable candidates\, thus mitigating the effect of noise on the properties of the inferred mod el. We review the empirical performance of a large number of cross-impact models\, comparing their strengths and weaknesses on a number of asset cla sses (futures\, stocks\, calendar spreads). Besides showing which models p erform better\, we argue that in presence of comparable statistical perfor mance\, which is often the case in a noisy world\, it is relevant to favor models that provide ex-ante theoretical guarantees on their behavior in l imit cases. From this perspective\, we advocate that the empirical validat ion of universal properties (symmetries\, invariances) should be regarded as holding a much deeper epistemological value than any measure of statist ical performance on specific model instances.\n\n//indico.ictp.it/event/93 50/ LOCATION:ZOOM MEETING Meeting ID: 475-819-702 URL://indico.ictp.it/event/9350/ END:VEVENT BEGIN:VEVENT SUMMARY:Optimal\, near-optimal\, and robust epidemic control DTSTART;VALUE=DATE-TIME:20200421T130000Z DTEND;VALUE=DATE-TIME:20200421T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9349@ictp.it DESCRIPTION:\nThe COVID-19 pandemic has highlighted the need for co ntrolmeasures that reduce the epidemic peak ("flattening the curve").\n Here we derive the optimal time-limited intervention for reducing peak epi demic prevalence in the standard Susceptible-Infectious-Recovered (SIR) mo del. We show that alternative\, more practical interventions can perform nearly as well as the provably optimal strategy. However\, none of these strategies are robust to implementation errors: mistiming the start of th e intervention by even a single week can be enormously costly\, for reali stic epidemic parameters. Sustained control measures\, though less effci ent than optimal and near-optimal time-limited interventions\, can be used in combination with time-limited strategies to mitigate the\ncatastrophic risks of mistiming.\n\n\n//indico.ictp.it/event/9349/ LOCATION:ZOOM MEETING Meeting ID: 475-819-702 URL://indico.ictp.it/event/9349/ END:VEVENT BEGIN:VEVENT SUMMARY:Population Dynamics of Emerging Diseases DTSTART;VALUE=DATE-TIME:20200508T130000Z DTEND;VALUE=DATE-TIME:20200508T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9363@ictp.it DESCRIPTION://indico.ictp.it/event/9363/ LOCATION:ZOOM MEETING Meeting ID: 475-819-702 URL://indico.ictp.it/event/9363/ END:VEVENT BEGIN:VEVENT SUMMARY:Monitoring of hemodynamics: hydrodynamic characteristics of blood flow and rheology of vessel tissue DTSTART;VALUE=DATE-TIME:20200520T090000Z DTEND;VALUE=DATE-TIME:20200520T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9376@ictp.it DESCRIPTION:\n Join Zoom Meeting : https://zoom.us/j/475.819702\n Meeting ID: 475-819-702 \n\n  \n\n Abstract:\n Cerebral vascular pathologies are a widespread and socially significant disease. In particular\, cerebral an eurysms can lead to stroke\, which ranks second in mortality in the world according to WHO. Large-scale studies conducted over the past decade in th e United States show that the existing criteria for making medical decisio ns (to perform surgery or to observe conservatively) have weak statistical significance\, which prompts many researchers to deal with this problem: preoperative modeling of the treatment of cerebral aneurysms.  At the sam e time\, the development of mathematical models\, as well as the creation of software complexes\, can be based only on a high-quality and large-scal e array of experimental data.In this study\, we consider two approaches th at we use to obtain personalized data on vascular pathology\, as well as m athematical and numerical modeling based on such experimental data.\n\n//i ndico.ictp.it/event/9376/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9376/ END:VEVENT BEGIN:VEVENT SUMMARY: Modelling Covid-19 in Brazil DTSTART;VALUE=DATE-TIME:20200527T130000Z DTEND;VALUE=DATE-TIME:20200527T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9379@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702Meeting ID: 4 75-819-702To obtain the password for this zoom meeting\, please contact ql s@ictp.it\n  \n\n Dr Kraenkel will present the current epidemic situation in Brazil and describe the efforts to contribute to public debate and dec ision making  by means mathematical modelling. Dr Kraenkel will talk abou t scenario exploration as a means of planning action and about indicators to follow the situation in almost real time.  Results refer to the munici pality  of São Paulo\, as well as for Brazil.\n\n//indico.ictp.it/event/ 9379/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9379/ END:VEVENT BEGIN:VEVENT SUMMARY:Statistical Learning Theory of geometrically structured data DTSTART;VALUE=DATE-TIME:20200609T130000Z DTEND;VALUE=DATE-TIME:20200609T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9381@ictp.it DESCRIPTION:Join Zoom Meeting https://zoom.us/j/475819702Meeting ID: 475- 819-702\n\n To obtain the PASSWORD for this zoom meeting\, please contact qls@ictp.it\n  \n\n The traditional approach of statistical physics to su pervised learning routinely assumes unrealistic generative models for the data: usually inputs are independent random variables\, uncorrelated with their labels. Only recently\, statistical physicists started to explore mo re complex forms of data\, such as equally-labelled points lying on (possi bly low dimensional) object manifolds. Here we provide a bridge between th is recently-established research area and the framework of statistical lea rning theory\, a branch of mathematics devoted to inference in machine lea rning. The overarching motivation is the inadequacy of the classic rigorou s results in explaining the remarkable generalization properties of deep l earning. We propose a way to integrate physical models of data into statis tical learning theory\, and address\, with both combinatorial and statisti cal mechanics methods\, the computation of the Vapnik-Chervonenkis entropy \, which counts the number of different binary classifications compatible with the loss class. As a proof of concept\, we focus on kernel machines a nd on two simple realizations of data structure introduced in recent physi cs literature: k-dimensional simplexes with prescribed geometric relation s and spherical manifolds (equivalent to margin classification). Entropy\, contrary to what happens for unstructured data\, is nonmonotonic in the s ample size\, in contrast with the rigorous bounds. Moreover\, data structu re induces a novel transition beyond the storage capacity\, which we advoc ate as a proxy of the nonmonotonicity\, and ultimately a cue of low genera lization error. The identification of a synaptic volume vanishing at the t ransition allows a quantification of the impact of data structure within r eplica theory\, applicable in cases where combinatorial methods are not av ailable\, as we demonstrate for margin learning.\n\n \n  \n\n\n//indico.i ctp.it/event/9381/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9381/ END:VEVENT BEGIN:VEVENT SUMMARY:Spatial Organization of Organisms and Functions in Bacterial Commu nities DTSTART;VALUE=DATE-TIME:20200616T130000Z DTEND;VALUE=DATE-TIME:20200616T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9399@ictp.it DESCRIPTION:Join Zoom Meeting https://zoom.us/j/475819702Meeting ID: 475- 819-702\n To obtain the PASSWORD for this zoom meeting\, please contact ql s@ictp.it\n Communities of interacting microbes perform fundamental proces ses on Earth\, such as cycling the elements and shaping our health. These processes arise from a dense network of interactions between individual c ells in the community. Many microbial communities are spatially structured \, thus interaction occur mostly between cells close in space. Therefore\, the spatial arrangement of different cell types determines the interactio ns that occur and therefore the processes that the whole community perform s. I observe microbial communities at the single-cell level and I reconstr uct the spatial network of interaction between cells.  I combine these ex periments with mathematical modeling to elucidate how community-level prop erties arise from the activity and the interactions between single cells.\ n\n//indico.ictp.it/event/9399/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9399/ END:VEVENT BEGIN:VEVENT SUMMARY:Simple equation describing brain hemodynamics DTSTART;VALUE=DATE-TIME:20200625T120000Z DTEND;VALUE=DATE-TIME:20200625T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9411@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702 \n \n \n Mee ting ID: 475-819-702\n\n If you haven't registered for previous QLS webina rs\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeti ng.\n\n\n The study of blood flow in the vicinity of cerebral vascular pat hologies is important both from an applied and from a scientific point of view. The main characteristics of blood flow are the dependence of velocit y and pressure on time. Currently\, the dependence of velocity on time can be measured by various methods (for example\, ultrasound). But\, currentl y non-invasive pressure measurement is impossible. \n\n In my talk I will present a mathematical model of blood flow\, which allows for a particular patient to restore the dependence of pressure on time by the dependence o f velocity on time. To build the model under study\, clinical data on the pressure and velocity of blood flow in the vessels of the brain in the vic inity of the pathology obtained during neurosurgical operations are used. The model is a nonlinear ordinary differential equation\, the coefficients of which are determined by clinical data. We study equation's properties as well as a correlation of equation properties and pathologies properties .\n\n\n\n//indico.ictp.it/event/9411/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9411/ END:VEVENT BEGIN:VEVENT SUMMARY:Collective Information Processing - Interplay of self-organization and function in collective (biological) systems DTSTART;VALUE=DATE-TIME:20200630T130000Z DTEND;VALUE=DATE-TIME:20200630T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9418@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702 \n Meeting I D: 475-819-702\n\n If you haven't registered for previous QLS webinars\, p lease contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n  \n\n Animal groups or cellular ensembles represent fascinating examples of self-organized biological systems. In contrast to non-living physical s ystems\, self-organized biological collectives are the result of long-term evolutionary adaptations to a specific ecological niche\, where collectiv e behavior provides evolutionary benefits to individual agents. However\, collective information processing is also always subject to constraints se t by the interaction mechanisms and corresponding self-organized dynamical structure. In my talk\, I will discuss examples from our recent research on how the adaptive structure of the collective may interplay with its bio logical function in terms of collective information processing.\n  \n\n\n \n//indico.ictp.it/event/9418/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9418/ END:VEVENT BEGIN:VEVENT SUMMARY:Entropy production\, work\, and thermodynamics of information unde r protocol constraints DTSTART;VALUE=DATE-TIME:20200714T143000Z DTEND;VALUE=DATE-TIME:20200714T153000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9426@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702 \n Meeting I D: 475-819-702\n\n If you haven't registered for previous QLS webinars\, p lease contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n  \n\n Abstract: Stochastic thermodynamics has derived various bounds on w ork and entropy production (EP) that apply to a broad class of nonequilibr ium processes. These bounds are typically achievable under idealized condi tions\, such as when arbitrary driving protocols can be employed. We deriv e bounds on the maximal extractable work and minimal EP which hold when th ere are constraints on the available driving protocols\, including constra ints on symmetries\, modularity\, and coarse-graining of the driving proto cols. We also show that our approach can be used to decompose information acquired in a measurement of a system into "useful information" (which can be used to extract work from the system) and “useless information” (w hich cannot be used to extract work from the system).\n\n\n\n//indico.ictp .it/event/9426/ LOCATION: URL://indico.ictp.it/event/9426/ END:VEVENT BEGIN:VEVENT SUMMARY:Active Processes in Biology driven by Molecular Motors DTSTART;VALUE=DATE-TIME:20200728T130000Z DTEND;VALUE=DATE-TIME:20200728T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9431@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702\n Meeting ID : 475-819-702If you haven't registered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n\n Acti ve processes are critical to maintain the proper function of all living ce lls. A large class of active processes is driven by molecular motors that transduce chemical energy into mechanical work. To understand their force- dependent dynamics\, precise measurements with optical traps have been est ablished. Because of the stochastic nature of the dynamics\, interpretatio n of the experimental data is not straightforward. We recently introduced a quantitative framework to analyze experimental data and to extract bioph ysical relevant quantities. We apply this method to data from human dynein to predict the maximal force it can generate.\n\n To further understand t he physical mechanisms underlying biological processes\, we investigate th e role of molecular motors in large-scale simulations. We integrate single motor behavior into such simulations to study force production during T-c ell activation.\n\n  \n\n  \n\n\n\n//indico.ictp.it/event/9431/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9431/ END:VEVENT BEGIN:VEVENT SUMMARY:First a war\, then a dance: Mechanochemical symmetry breaking in s ensory organ development DTSTART;VALUE=DATE-TIME:20200922T140000Z DTEND;VALUE=DATE-TIME:20200922T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9452@ictp.it DESCRIPTION:https://zoom.us/j/475819702\nMeeting ID: 475-819-702\nIf you h aven't registered for previous QLS webinars\, please contact qls@ictp.it t o obtain the PASSWORD for this zoom meeting.\nActively regulated symmetry breaking\, which is ubiquitous in biological cells\, underlies phenomena s uch as directed cellular movement and morphological polarization. Here\, w e investigate how an organ-level polarity pattern emerges through symmetry breaking at the cellular level during the formation of a mechanosensory o rgan. Combining theory\, genetic perturbations and in vivo imaging\, we st udy the development and regeneration of the fluid-motion sensors in the ze brafish’s lateral line. We find that two interacting symmetry-breaking e vents—one mediated by biochemical signalling and the other by cellular m echanics—give rise to precise rotations of cell pairs\, which produce a mirror-symmetric polarity pattern in the receptor organ.\n\n//indico.ictp. it/event/9452/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9452/ END:VEVENT BEGIN:VEVENT SUMMARY:Thermodynamics of error correction DTSTART;VALUE=DATE-TIME:20200929T120000Z DTEND;VALUE=DATE-TIME:20200929T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9453@ictp.it DESCRIPTION:https://zoom.us/j/475819702\n\nMeeting ID: 475-819-702\n\nIf y ou haven't registered for previous QLS webinars\, please contact qls@ictp. it to obtain the PASSWORD for this zoom meeting.\n\n\nBiological systems a re able to replicate information with outstanding accuracy. In biochemical pathways such as those leading to DNA duplication or protein translations \, different monomers can be distinguished because of equilibrium free-ene rgy differences or via non-equilibrium mechanisms. I will show how\, in si mple copying reactions\, these two discrimination modes are mutually exclu sive and lead to opposite tradeoffs between error\, dissipation and reacti on velocity. In multi-step reactions\, such as in kinetic proofreading\, these different modes can be combined to improve overall accuracy. In the second part of my talk\, I will present recent results about fluctuations in accuracy and speed of these reactions.\n\n//indico.ictp.it/event/9453/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9453/ END:VEVENT BEGIN:VEVENT SUMMARY:Physical laws of unsupervised learning in neural network DTSTART;VALUE=DATE-TIME:20201006T130000Z DTEND;VALUE=DATE-TIME:20201006T150000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9456@ictp.it DESCRIPTION:https://zoom.us/j/475819702\n\nMeeting ID: 475-819-702\n\nIf y ou haven't registered for previous QLS webinars\, please contact qls@ictp. it to obtain the PASSWORD for this zoom meeting.\n\n\nAbstract:\nWe provid e a complete picture about mechanisms of unsupervised learning\, i.e.\, in a simple neural network system\, unsupervised learning can be interpreted as breaking a series of symmetries\, driven by increasing observations. F irst\, it is the spontaneous symmetry breaking starting the concept format ion\, then the permutation symmetry among hidden units is spontaneously br oken\, including in sequence two types: the student side is the first\, an d the teacher (planted truth) is the second. In this paper\, we also analy tically prove that the learning threshold in a simple neural network that triggers a spontaneous symmetry breaking (concept-formation) does not depe nd on the number of hidden neurons (here two for a minimal model) once thi s number is finite\, for the correlation-free case. The underlying physics is the partition function factorization for which we give a proof. Moreov er\, our analytical result reveals that the weak correlation among recepti ve fields of hidden neurons significantly reduces the learning threshold\, which is consistent with the non-redundant weight assumption popular in s ystem neuroscience and machine learning! By studying the minimal model\, w e are excited to reveal the inner workings of unsupervised learning\, a fu ndamental process governing artificial and biological intelligence. We exp ect this work may open doors towards physical laws governing neural learni ng.\n\n//indico.ictp.it/event/9456/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9456/ END:VEVENT BEGIN:VEVENT SUMMARY:ICTP QLS-AP Colloquium AI / How TinyML Could Help Developing Count ries DTSTART;VALUE=DATE-TIME:20201013T140000Z DTEND;VALUE=DATE-TIME:20201013T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9460@ictp.it DESCRIPTION:Zoom webinar. When: Oct 13\, 2020 04:00 PM Rome Topic: AI coll oquia Register in advance for this webinar:https://zoom.us/webinar/registe r/WN_qidXVl5URkKbG5kJUAMB0A\nAfter registering\, you will receive a confir mation email containing information about joining the webinar.\nShould you not be able to join the Webinar\, the Colloquium is also available in liv e streaming at:http://ictp.it/livestream\nAbstract\nMachine learning on em bedded hardware is a new field that's only beginning to be used in the dev eloped world\, but it has some characteristics that may make it particular ly appropriate for developing countries. The hardware and software require d to train students\, advance research\, and build products is much cheape r than the equivalents for Cloud ML\, and since the devices don't need rel iable network or even mains power connections\, they can be deployed almos t anywhere. Pete Warden is the Technical Lead of the TensorFlow Micro open source project from Google\, and has previously helped projects like Plan t Village use on-device deep learning in West Africa. In this talk he'll i ntroduce some of the possibilities that edge compute opens up for the worl d outside the West\, and will be looking to learn from the audience what p roblems in their domains it might be applicable to.\n\n//indico.ictp.it/ev ent/9460/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9460/ END:VEVENT BEGIN:VEVENT SUMMARY:How do microbial communities respond to perturbations? DTSTART;VALUE=DATE-TIME:20201020T130000Z DTEND;VALUE=DATE-TIME:20201020T150000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9457@ictp.it DESCRIPTION:https://zoom.us/j/475819702\n\nMeeting ID: 475-819-702\n\nIf y ou haven't registered for previous QLS webinars\, please contact qls@ictp. it to obtain the PASSWORD for this zoom meeting.\n\nAbstract: \nMicrobial communities play myriad roles in ecosystems\, from fixing atmospheric nitr ogen\, to recycling organic matter to promoting host health. In many cases \, e.g. in the human gut ecosystem\, these communities display alternative stable states associated to health and dysbiosis. Not only this\, but our daily life commonly gives rise to perturbations (changes in diet\, infect ions\, or exposure to antibiotics) that can induce community shifts toward s alternative stable states. However\, the mechanisms driving these shifts are\, in general\, poorly understood. In this talk\, I will introduce an experimental model system that I use to interrogate\, quantitatively\, mic robial community resilience to perturbations and transitions between stabl e states. I will show that the arrival of an invader species can often tri gger transitions between stable states of microbial communities\, even if the invader itself does not survive the transition. I will also describe o ngoing work in which I expose this model community to antibiotic perturbat ions. In this case\, our theoretical model predicts\, and experiments conf irmed\, that ecological parameters such as growth and migration rates can be more important in determining the response of the community than the sp ecific kind of antibiotics that is being used.\n\n//indico.ictp.it/event/9 457/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9457/ END:VEVENT BEGIN:VEVENT SUMMARY:Dissecting liquid-liquid phase separation in engineered and biolog ical systems DTSTART;VALUE=DATE-TIME:20201027T150000Z DTEND;VALUE=DATE-TIME:20201027T170000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9470@ictp.it DESCRIPTION:https://zoom.us/j/475819702 \nMeeting ID: 475-819-702\nIf you haven't registered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\nAbstract:Liquid-liquid phas e separation (LLPS) of biopolymers recently emerged as a ubiquitous drivin g force for intracellular self-assembly. The resulting LLPS dynamics appea r to orchestrate wide-ranging cellular mechanisms\, from genome organizati on to cell division. Despite remarkable progress\, studying intracellular LLPS remains challenging. In this talk\, I will present our recent progres s towards dissecting the LLPS of intrinsically-disordered proteins (IDPs) and its biological significance. I will highlight our two main approaches: (1) the engineering of lower and upper critical solution temperature (LCS T and UCST\, respectively) LLPS behavior in IDPs (Nature Materials 2015\, Science Advances 2019)\, and (2) the probing of intracellular LLPS dynamic s in mammalian skin (Science 2020). Dwelling on skin\, I will explain how the biophysical lens of LLPS sheds new light into the enigmatic process of skin barrier formation. Using this system\, I will also describe novel bi omolecular tools to probe LLPS in tissue biology. Finally\, I will discuss engineered IDPs that exhibit non-equilibrium LLPS. These behaviors differ from the established biophysical frameworks to study LLPS of IDPs\, which were inspired by work on synthetic polymers that exhibit equilibrium LLPS . Overall\, my goal is to inspire students and scientists at ICTP to conne ct with the research frontier of cellular mechanisms driven by LLPS. This nascent field offers fruitful ground for exploration of experimental and t heoretical biophysical approaches. \n\n\n\n//indico.ictp.it/event/9470/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9470/ END:VEVENT BEGIN:VEVENT SUMMARY:Symmetry breaking during morphogenesis of a mechanosensory organ DTSTART;VALUE=DATE-TIME:20201103T150000Z DTEND;VALUE=DATE-TIME:20201103T170000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9482@ictp.it DESCRIPTION:https://zoom.us/j/475819702 \nMeeting ID: 475-819-702\nIf you haven't registered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\nAbstract:\nThe development of mechanosensory epithelia\, such as those of the auditory and vestibular systems\, results in the precise orientation of mechanosensory hair cells and consequently directional sensitivity. After division of a precursor c ell in the zebrafish's lateral line\, the daughter hair cells differentiat e with opposite mechanical sensitivity. This process produces neuromasts c ontaining equal numbers of hair cells of two opposite polarities\, half of them sensitive to caudal water movement and half to rostral flow.\nThroug h a combination of theoretical and experimental approaches\, we show that Notch-mediated lateral inhibition produces a bistable switch that reliably gives rise to hair cell pairs of opposite polarity. Using a mathematical model of the process\, we predict the outcome of several genetic and chemi cal alterations to the system\, which we then confirm experimentally.\n\n/ /indico.ictp.it/event/9482/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9482/ END:VEVENT BEGIN:VEVENT SUMMARY:Morphogenesis\, mechanics\, and malignancy: How complex tissue arc hitectures arise in cancers DTSTART;VALUE=DATE-TIME:20201110T150000Z DTEND;VALUE=DATE-TIME:20201110T170000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9486@ictp.it DESCRIPTION:https://zoom.us/j/475819702\nMeeting ID: 475-819-702\nIf you h aven't registered for previous QLS webinars\, please contact qls@ictp.it t o obtain the PASSWORD for this zoom meeting.\nAbstract:\nMorphogenesis\, m echanics\, and malignancy: How complex tissue architectures arise in cance rs\nTissues are precisely built in their three-dimensional shape and the a rrangement of specific cell types. This architecture breaks down as cancer -causing mutations transform normal tissue into tumor. How these architect ural changes occur – through biophysical forces and gene regulatory netw orks that must coordinate across length and time scales – is a fundament al question in developmental and cancer biology. In this talk\, I will dis cuss recently discovered mechanisms of mammalian tumor morphogenesis. Usin g a combination of computational modeling\, genetic manipulations and biop hysical measurements in vivo\, I will describe how the biophysical drivers of tumor morphogenesis are distinct between simple monolayer and multilay ered epithelia\, and how initial architecture of the tissue-of-origin infl uences tumor morphogenesis. Our findings suggest that mechanical forces su rrounding the proliferative cells within the tumor\, so-called ‘cancer s tem cells’\, shape architecture of premalignant tumors and ultimately in fluence their progression into malignant cancer.\n\n//indico.ictp.it/event /9486/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9486/ END:VEVENT BEGIN:VEVENT SUMMARY:What are we measuring with magnetic resonance spectroscopy: A neur al field model of GABA/Glu synaptic transmission DTSTART;VALUE=DATE-TIME:20201117T140000Z DTEND;VALUE=DATE-TIME:20201117T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9487@ictp.it DESCRIPTION:https://zoom.us/j/475819702\nMeeting ID: 475-819-702\nIf you h aven't registered for previous QLS webinars\, please contact qls@ictp.it t o obtain the PASSWORD for this zoom meeting.\nAbstract: Long-term changes to synaptic plasticity due to transcranial direct current stimulation (tDC S) have been linked to the dynamics of glutamate and GABA via the quantifi cation of the neurotransmitter’s concentration from 1H-NMR spectroscopy (MRS) in vivo in a single scan session. However\, our understanding of pre cisely which physiological processes underlie changes to GABA/Glu concentr ations in response to stimulation remains unclear. In other words\, what a re we measuring with MRS in these situations? One theory is that we are me asuring compartmental shifts of neurotransmitters from pre-synaptic vesicl es to more visible synaptic\, extracellular and cytosolic pools. It is imp ortant to understand because there are currently various therapeutic techn iques used to target and modulate neurochemical distributions with the aim of promoting functional recovery (from stroke\, for example). In this tal k\, I present the development of a neural model that incorporates neurotra nsmitter dynamics and calcium-dependent plasticity to allow us to probe th e relationship between stimulation\, neurotransmitter activity\, and plast icity. I simulate tDCS to drive fluctuations in the model\, and in so doin g I demonstrate how tDCS modifies plasticity as well as the similarities b etween the model prediction and empirical MRS data. This modelling approac h will allow mechanistically informed optimization of current stimulation parameters.\n \n\n//indico.ictp.it/event/9487/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9487/ END:VEVENT BEGIN:VEVENT SUMMARY:Machine learning the thermodynamic arrow of time DTSTART;VALUE=DATE-TIME:20210112T150000Z DTEND;VALUE=DATE-TIME:20210112T170000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9520@ictp.it DESCRIPTION:https://zoom.us/j/475819702 \nMeeting ID: 475-819-702\nIf you haven't registered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n Abstract: The asymmetry in the flow of events that is expressed by the phrase ‘time’s arrow’ t races back to the second law of thermodynamics. In the microscopic regime\ , fluctuations prevent us from discerning the direction of time’s arrow with certainty. Here\, we show that a machine learning algorithm that is t rained to infer the direction of time’s arrow identifies entropy product ion as the relevant physical quantity in its decision-making process. Effe ctively\, the algorithm rediscovers the fluctuation theorem as the underly ing thermodynamic principle. Our results indicate that machine learning te chniques can be used to study systems that are out of equilibrium\, and ul timately to answer open questions and uncover physical principles in therm odynamics.\n\n//indico.ictp.it/event/9520/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9520/ END:VEVENT BEGIN:VEVENT SUMMARY:Dynamical processes in cells and tissues DTSTART;VALUE=DATE-TIME:20210122T150000Z DTEND;VALUE=DATE-TIME:20210122T170000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9525@ictp.it DESCRIPTION:https://zoom.us/j/475819702\nMeeting ID: 475-819-702\nIf you h aven't registered for previous QLS webinars\, please contact qls@ictp.it t o obtain the PASSWORD for this zoom meeting.\nAbstract: I will present two examples of dynamics and self-organization in living systems. Differentia tion of adult stem cells into specialised cells plays a very important rol e in tissue regeneration and wound healing. Bone healing is a highly compl ex and tightly regulated physiological process that involves proliferation of human mesenchymal stem cells and osteogenic differentiation of human m esenchymal stem cells into osteoblasts (bone cells). In vitro osteoinducti on via external electric field is a promising approach to build implants f or bone regeneration. I will present a general mean-field theoretical fram ework that takes into account processes at the scale of a single cell and describes the dynamics of a stem cell population [1]. The theory is tested with in vitro experiments involving electrical stimulation of human mesen chymal stem cells. The analysis shows that proliferation rate is cell dens ity-dependent and that electric field influences only osteogenic different iation. At the tissue level\, leaf vein morphogenesis and patterning invol ves cell shape changes \,and\, is under the regulation of plant growth hor mone auxin. How auxin coordinates the developmental regulation of leaf vei n is not well understood. I will present a cell based model that captures the interplay between key biochemical processes and cell mechanics and des cribes early stages of mid-vein development in a growing leaf primordia [2 ]. Comparison of the model with experiments shows that the formation of ea rly vein patterns may essentially be described in terms of auxin biosynthe sis\, its transport\, and growth-dependent mechanical feedback from surrou nding tissues.\n [1] A general theoretical framework to study the influenc e of electrical fields on mesenchymal stem cells
\n\n Dawson\, Lee\, van Rienen\, Appali\, Frontiers Bioengineering and Biotechnology (2020) doi: https://doi.org/10.3389/fbioe.2020.557447\n\n [2] Auxin biosynthesis and c ellular efflux act together to regulate leaf vein patterning\n\n Kneuper\, Teale\, Dawson\, Tsugeki\, Katifori\, Palme\, Ditengou\, Journal of Exper imental Botany (2020) doi: https://doi.org/10.1093/jxb/eraa501\n\n\n\n//in dico.ictp.it/event/9525/ LOCATION:zoom URL://indico.ictp.it/event/9525/ END:VEVENT BEGIN:VEVENT SUMMARY:The energetic cost of biological waves DTSTART;VALUE=DATE-TIME:20210126T140000Z DTEND;VALUE=DATE-TIME:20210126T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9543@ictp.it DESCRIPTION:\n https://zoom.us/j/475819702 \n Meeting ID: 475-819-702\n If you haven't registered for previous QLS webinars\, please contact qls@ict p.it to obtain the PASSWORD for this zoom meeting.Abstract: Living matter consumes energy at the microscopic scale to drive emergent\, large-scale b ehaviors such as traveling waves or coherent oscillations. These complex\, spatiotemporal phenomena play pivotal roles in essential life processes\, such as cell division\, cell polarization\, and collective migration. Whi le much work has focused on understanding the dynamics of these behaviors\ , we lack a thorough understanding of their energetic underpinnings. To ad dress this\, we introduce a measure of irreversibility we term the entropy production factor that estimates energy dissipation across time and space for spatially extended systems. We use our method to study the emergence of synchronized oscillations in a model biochemical oscillator\, finding s ignatures of the onset of oscillations in the total irreversibility\, whil e signatures of synchronization are found in the distribution of irreversi bility across spatiotemporal frequencies. This work provides a tool for un derstanding the connection between microscopic energy consumption and macr oscopic dynamic phenomena in non-equilibrium systems.\n\n\n\n//indico.ictp .it/event/9543/ LOCATION:zoom URL://indico.ictp.it/event/9543/ END:VEVENT BEGIN:VEVENT SUMMARY:Quantifying dynamic stability and signal propagation: Lyapunov spe ctra of recurrent neural networks DTSTART;VALUE=DATE-TIME:20210202T140000Z DTEND;VALUE=DATE-TIME:20210202T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9550@ictp.it DESCRIPTION:\n https://zoom.us/j/475819702\n Meeting ID: 475-819-702\n If you haven't registered for previous QLS webinars\, please contact qls@ictp .it to obtain the PASSWORD for this zoom meeting.\n Abstract: Brains proce ss information through the collective dynamics of large neural networks. C ollective chaos was suggested to underlie the complex ongoing dynamics obs erved in cerebral cortical circuits and determine the impact and processin g of incoming information streams. While dynamic mean-field theory has unc overed key properties of recurrent network models such as the onset of cha os and their largest Lyapunov exponent\, fundamental features of their dyn amics remain unknown. In particular\, chaotic dynamics in dissipative high -dimensional systems takes place on a subset of phase space of reduced dim ension and is organized by a complex tangle of stable\, neutral and unstab le manifolds. Key topological invariants of this phase space structure suc h as attractor dimension\, and Kolmogorov-Sinai entropy so far remained el usive.\n Here we calculate the complete Lyapunov spectrum of recurrent neu ral networks. We show that chaos in these networks is extensive with a siz e-invariant Lyapunov spectrum and characterized by attractor dimensions mu ch smaller than the number of phase space dimensions. The attractor dimens ion and entropy rate increases with coupling strength near the onset of ch aos but decreases far from onset\, reflecting a reduction in the number of unstable directions. We find that near the onset of chaos\, for very inte nse chaos\, and discrete-time dynamics\, random matrix theory provides goo d analytical approximations to the full Lyapunov spectrum. We show that a generalized time-reversal symmetry of the network dynamics induces a point -symmetry of the Lyapunov spectrum reminiscent of the symplectic structure of chaotic Hamiltonian systems. Temporally fluctuating input can drastica lly reduce both the entropy rate and the attractor dimension. For trained recurrent networks\, we find that Lyapunov spectrum analysis provides a qu antification of error propagation and stability achieved by distinct learn ing algorithms. Our methods apply to systems of arbitrary connectivity\, a nd we describe a comprehensive set of controls for the accuracy and conver gence of Lyapunov exponents.\n Our results open a novel avenue for charact erizing the complex dynamics of recurrent neural networks and the geometry of the corresponding high-dimensional chaotic attractor. They also highli ght the potential of Lyapunov spectrum analysis as a diagnostic for machin e learning applications of recurrent networks.\n\n\n\n//indico.ictp.it/eve nt/9550/ LOCATION:zoom URL://indico.ictp.it/event/9550/ END:VEVENT BEGIN:VEVENT SUMMARY:The impact of data structure on learning in two-layer neural netwo rks (Lecture 1) DTSTART;VALUE=DATE-TIME:20210209T140000Z DTEND;VALUE=DATE-TIME:20210209T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9563@ictp.it DESCRIPTION:https://zoom.us/j/475819702\nMeeting ID: 475-819-702\nIf you h aven't registered for previous QLS webinars\, please contact qls@ictp.it t o obtain the PASSWORD for this zoom meeting.\nAbstract: Understanding the impact of data structure on learning in neural networks remains a key chal lenge for the theory of neural networks.\nIn these two lectures\, we will discuss how to go beyond the simple i.i.d. modelling paradigm in the teac her-student setup by studying neural networks trained on data drawn from s tructured generative models. \nOur discussion will center around two resu lts:\n(1) We give rigorous conditions under which a class of generative models shares key statistical properties with an appropriately chosen Gaus sian feature model.\n(2) We use this Gaussian equivalence to analyse the d ynamics of two-layer neural networks trained using one-pass stochastic gra dient descent on data drawn from a large class of generators.  \n \nI will try to make these lectures self-contained. \nThey will be mostly bas ed on the following two papers:\n \n[1] Goldt\, S.\, Mézard\, M.\, Krza kala\, F. and Zdeborová\, L.\, 2020.\n     Modeling the Influence of D ata Structure on Learning in Neural Networks: The Hidden Manifold Model.  Physical Review X\, 10(4)\, p.041044.\n     https://arxiv.org/abs/1909 .11500\n[2] Goldt\, S.\, Reeves\, G.\, Loureiro\, B.\, Mézard\, M.\, Krz akala\, F. and Zdeborová\, L.\, 2020.\n     The Gaussian equivalence o f generative models for learning with two-layer neural networks.       under review\; https://arXiv.org/abs/2006.14709\n\n//indico.ictp.it/event/ 9563/ LOCATION:Zoom meeting URL://indico.ictp.it/event/9563/ END:VEVENT BEGIN:VEVENT SUMMARY:"The impact of data structure on learning in two-layer neural netw orks" (Lecture 2) DTSTART;VALUE=DATE-TIME:20210216T140000Z DTEND;VALUE=DATE-TIME:20210216T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9578@ictp.it DESCRIPTION:https://zoom.us/j/475819702\nMeeting ID: 475-819-702\nIf you h aven't registered for previous QLS webinars\, please contact qls@ictp.it t o obtain the PASSWORD for this zoom meeting.\n \nAbstract: Understanding the impact of data structure on learning in neural networks remains a key challenge for the theory of neural networks.\nIn these two lectures\, we w ill discuss how to go beyond the simple i.i.d. modelling paradigm in the teacher-student setup by studying neural networks trained on data drawn fr om structured generative models. \nOur discussion will center around two results:\n(1) We give rigorous conditions under which a class of generat ive models shares key statistical properties with an appropriately chosen Gaussian feature model.\n(2) We use this Gaussian equivalence to analyse t he dynamics of two-layer neural networks trained using one-pass stochastic gradient descent on data drawn from a large class of generators.   \n I will try to make these lectures self-contained. \nThey will be mostly b ased on the following two papers: \n[1] Goldt\, S.\, Mézard\, M.\, Krza kala\, F. and Zdeborová\, L.\, 2020.\n     Modeling the Influence of D ata Structure on Learning in Neural Networks: The Hidden Manifold Model.  Physical Review X\, 10(4)\, p.041044.\n     https://arxiv.org/abs/1909 .11500\n[2] Goldt\, S.\, Reeves\, G.\, Loureiro\, B.\, Mézard\, M.\, Krz akala\, F. and Zdeborová\, L.\, 2020.\n     The Gaussian equivalence o f generative models for learning with two-layer neural networks.       under review\; https://arXiv.org/abs/2006.14709\n\n//indico.ictp.it/event/ 9578/ LOCATION:zoom meeting URL://indico.ictp.it/event/9578/ END:VEVENT BEGIN:VEVENT SUMMARY:Applications of Random Matrix Theory to single-cell biology DTSTART;VALUE=DATE-TIME:20210223T140000Z DTEND;VALUE=DATE-TIME:20210223T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9588@ictp.it DESCRIPTION:Zoom Meeting ID to attend the online seminar: 475-819-702\nJoi n Zoom Meeting:https://zoom.us/j/475819702\nIf you haven't registered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meetingAbstract: Recent advances in single-cell technologies provide a unique opportunity to study tumor heterogeneity\, which is one of the main problems in developing therapies for cancer patients. Associat ed technological and biological problems such as sparsity of the data and stochasticity of gene expression provide fundamental challenges for rigoro us subpopulation identification. Random Matrix Theory (RMT) allows to deal with those challenges in a systematic way thanks to the universality prop erties at the eigenvalue and eigenvector levels . Synergetic use with mach ine learning algorithms provides a unique avenue to distinguish single-cel l heterogeneous subpopulations. I will present an introduction to the appl ications of RMT in single-cell biology and an application to construct a c ell atlas of the human and mouse prostate epithelium\n\n//indico.ictp.it/e vent/9588/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9588/ END:VEVENT BEGIN:VEVENT SUMMARY:The deep Boltzmann machine on the Nishimori line DTSTART;VALUE=DATE-TIME:20210302T153000Z DTEND;VALUE=DATE-TIME:20210302T173000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9594@ictp.it DESCRIPTION:https://zoom.us/j/475819702\nMeeting ID: 475-819-702\nIf you h aven't registered for previous QLS webinars\, please contact qls@ictp.it t o obtain the PASSWORD for this zoom meeting.\n Abstract: In this talk I wi ll address the direct problem of the deep (restricted) Boltzmann machine ( DBM) in the form of a multi-partite spin-glass model. Multi-partite models emerge as a natural generalization of mean field models where the permuta tion symmetry among the spins is weakened and preserved only inside each p artition of the system. These models were studied both in deterministic an d disordered settings and in general their solution requires a technical c onvexity assumption on the interactions that does not hold for the DBM. Th is feature makes it a challenging problem. Indeed\, the DBM is still unsol ved for centered interactions. On the contrary\, in my talk I will show ho w to exactly and rigorously compute its thermodynamic limit on the Nishimo ri line\, a particular subregion of the phase space where replica symmetry is proved to hold. As I will try to explain\, besides its interest in Sta tistical Mechanics the Nishimori line also provides a bridge between disor dered systems and Statistical Inference. Finally\, I will discuss how the geometry of the model\, especially the relative sizes of the layers\, can influence the phase transition. This is expected to provide some insight o n the optimal shape of a deep architecture in order to have a non trivial spontaneous magnetization.\n\n//indico.ictp.it/event/9594/ LOCATION:zoom meeting URL://indico.ictp.it/event/9594/ END:VEVENT BEGIN:VEVENT SUMMARY:Theory of gating in recurrent neural networks DTSTART;VALUE=DATE-TIME:20210309T140000Z DTEND;VALUE=DATE-TIME:20210309T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9600@ictp.it DESCRIPTION:Zoom Meeting ID to attend the online seminar: 475-819-702\nJoi n Zoom Meeting: https://zoom.us/j/475819702\nIf you haven't registered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n Abstract: Recurrent neural networks (RNNs) are p owerful dynamical models\, widely used in machine learning (ML) for proces sing sequential data\, and in neuroscience\, to understand the emergent p roperties of networks of real neurons. Prior theoretical work in understa nding the properties of RNNs has focused on networks with additive inter actions. However\, gating – i.e. multiplicative – interactions are ubi quitous in real neurons\, and gating is also the central feature of the be st-performing RNNs in ML. Here\, we study the consequences of gating for the dynamical behavior of RNNs. We show that gating leads to slow modes a nd a novel\, marginally-stable state. The network in this marginally-sta ble state can function as a robust integrator\, and unlike previous approa ches\, gating permits this function without parameter fine-tuning or spe cial symmetries. We study the long-time behavior of the gated network usi ng its Lyapunov spectrum\, and provide a novel relation between the maxi mum Lyapunov exponent and the relaxation time of the dynamics. Gating is a lso shown to give rise to a novel\, discontinuous transition to chaos\, where the proliferation of critical points (topological complexity) is de coupled from the appearance of chaotic dynamics (dynamical complexity)\, in contrast to a seminal result for additive RNNs. The rich dynamical beha vior is summarized in a phase diagram indicating critical surfaces and r egions of marginal stability – thus\, providing a map for principled pa rameter choices to ML practitioners. Finally\, we develop a field theory f or gradients that arise in training\, by combining the adjoint formalism from control theory with the dynamical mean-field theory. This paves the way for the use of powerful field theoretic techniques to study trainin g and gradients in large RNNs.\n\n//indico.ictp.it/event/9600/ LOCATION:zoom URL://indico.ictp.it/event/9600/ END:VEVENT BEGIN:VEVENT SUMMARY:Mathematical modeling of belowground plant interactions DTSTART;VALUE=DATE-TIME:20210323T140000Z DTEND;VALUE=DATE-TIME:20210323T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9614@ictp.it DESCRIPTION:Zoom Meeting ID to attend the online seminar: 475-819-702\nJoi n Zoom Meeting:https://zoom.us/j/475819702\nIf you haven't registered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n Abstract: Plant roots determine carbon uptake\, s urvivorship\, and agricultural yield and represent a large proportion of t he world’s vegetation carbon pool. The study of belowground competition\ , unlike aboveground shoot competition\, is hampered by our inability to o bserve roots. We have few observations of intact root systems in soil and lack a comprehensive theory for root system responses to their environment .\n In this presentation\, I will first review previous efforts to explain plant belowground competition and discuss how they lead to seemingly cont radictory predictions. Then\, I will introduce our recent work and show ho w it resolves existing controversy and provides a unifying framework to st udy belowground plant interactions. I will conclude by discussing future r esearch lines that depart from our results and how they can be addressed w ith extensions of our original model\n  \n\n//indico.ictp.it/event/9614/ LOCATION:zoom seminar URL://indico.ictp.it/event/9614/ END:VEVENT BEGIN:VEVENT SUMMARY:On generalized information measures for time-varying transmission channels DTSTART;VALUE=DATE-TIME:20210406T130000Z DTEND;VALUE=DATE-TIME:20210406T150000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9624@ictp.it DESCRIPTION:https://zoom.us/j/475819702\nMeeting ID: 475-819-702\nIf you h aven't registered for previous QLS webinars\, please contact qls@ictp.it t o obtain the PASSWORD for this zoom meeting.\n \nThe abstract will be pub lished as soon as possible\n\n//indico.ictp.it/event/9624/ LOCATION:zoom meeting URL://indico.ictp.it/event/9624/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Unveiling steady states in continuous cell cultures : A two way tale DTSTART;VALUE=DATE-TIME:20210413T130000Z DTEND;VALUE=DATE-TIME:20210413T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9629@ictp.it DESCRIPTION:\nJoin Jitsi meeting: meet.jit.si/ICTP-QLS-Seminar\n \n  \n\ nPlease note that not all internet browsers are supported (e.g. Safari)\, and Chrome is preferable\nIf you haven't registered for previous QLS webin ars\, please contact qls@ictp.it to obtain the PASSWORD for this meeting.A bstract:\n\n We first present a model for continuous cell culture coupling the intra-cellular metabolism of the cells to the external variables desc ribing the state of the bioreactor. We show that the ratio between the ce ll density and the dilution rate is an ideal control parameter to fix a st eady state with desired metabolic properties and that toxic byproduct accu mulation may drive the appearance of multiple steady states. Besides when the heterogeneity in the culture is taken into account the metabolic lands cape turns more complex including multiple metabolic switches. Then\, the se results\, combined with the Maximum Entropy principle\, were exploited to solve the inverse problem\, i.e.\, given the dilution rate and the cell density in a chemostat to infer the consumption and production of specifi c metabolites in the culture. We test the consistency of our results in a toy model of a metabolic network and in real experimental data and show th at our approach to infer metabolite concentration outperforms standards te chniques in the literature.\n  \n\n//indico.ictp.it/event/9629/ LOCATION:meet.jit.si/ICTP-QLS-Seminar URL://indico.ictp.it/event/9629/ END:VEVENT BEGIN:VEVENT SUMMARY:Gambling\, thermodynamic uncertainty and bet-hedging DTSTART;VALUE=DATE-TIME:20210420T130000Z DTEND;VALUE=DATE-TIME:20210420T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9649@ictp.it DESCRIPTION:Join Zoom Meeting https://zoom.us/j/475819702\nMeeting ID: 47 5-819-702If you haven't registered for previous QLS webinars\, please cont act qls@ictp.it to obtain the PASSWORD for this zoom meeting\n Abstract:\n \n \n Kelly's solution to allocating bets in a horse race\, or other gamb ling games\, maximizes the long-term growth rate\, but is known to be risk y. Here\, I will describe optimal betting strategies that give the highest capital growth rate while keeping a certain low value of risk\, that is\, an optimal trade-off between the average and the fluctuations. This trade -off is also embodied in a general bound similar to thermodynamic uncertai nty relations. Finally\, I will discuss possible future applications in bi ology. This trade-off is also believed to operate in situations known gene rically in biology as "bet-hedging". Organisms may exchange a reduction of their present fitness for increased fitness in a possible future stressfu l condition. Some type of bet-hedging mechanism is believed to operate in the production of seeds by plants or in the resistance to antibiotics.\n\n \n\n//indico.ictp.it/event/9649/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9649/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Rulers\, timers and transport for length control in l ong cell protrusions DTSTART;VALUE=DATE-TIME:20210427T130000Z DTEND;VALUE=DATE-TIME:20210427T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9651@ictp.it DESCRIPTION:Join Zoom Meeting https://zoom.us/j/475819702\nMeeting ID: 47 5-819-702\nIf you haven't registered for previous QLS webinars\, please co ntact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\nAbstract: \nA living cell uses its long cylindrical appendages for locomotion and se nsory purposes. Therefore\, assembling and maintaining a protrusion of cor rect length is crucial for its survival and overall performance. Usually t he protrusions lack the machinery for the synthesis of building blocks and imports them from the cell body.\nWhat are the unique features of the tra nsport logistics which facilitate the exchange of these building blocks be tween the cell and the protrusion? What kind of ‘rulers’ and ‘timers ’ does the cell use for constructing its appendages of correct length? H ow do the multiple appendages coordinate and communicate among themselves during different stages of their existence? How frequently do the fluctuat ions drive the length of these dynamic protrusions out of the acceptable b ounds?\nIn this talk\, I'll present an overview of our theoretical models that address these questions in the context of eukaryotic flagella.\n\n//i ndico.ictp.it/event/9651/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9651/ END:VEVENT BEGIN:VEVENT SUMMARY:Joint ICTP-SISSA-UniTS Data Science Seminar - Single-cell dynamics in chromatin regulation: machine learning methods to understand gene regu lation DTSTART;VALUE=DATE-TIME:20210504T130000Z DTEND;VALUE=DATE-TIME:20210504T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9653@ictp.it DESCRIPTION:The ICTP Quantitative Life Sciences Section is glad to announc e the Joint ICTP-SISSA-UniTS Data Science Seminar by Jake Yeung\, Human Fr ontier Science Program (HFSP) Fellow\, Hubrecht Institute for Developmenta l Biology and Stem Cell Research\, Utrecht\, The Netherlands\n \n Abstract :\n Cells are the basic units of life. Although every cell in the body has essentially the same genetic code\, histone modifications decorate the ge nome of different cells to establish distinct regions of the chromatin tha t are active or repressed. These cell type-specific chromatin states allow cells to turn on different genes and perform different functions. Recentl y\, new techniques have begun to map these histone modifications at the si ngle-cell level\, opening up new machine learning opportunities to infer g ene regulatory principles at unprecedented resolutions.\n  \n In this tal k\, I present two projects highlighting machine learning methods applied t o these experimental technologies to reveal global insights in chromatin r egulation in single cells.\n  \n First\, I present a new technology\, sor tChIC (sort-assisted chromatin immunocleavage)\, and apply it to understan d chromatin regulation during blood formation. I show machine learning met hods to learn interpretable maps from these high-dimensional count data to predict regulators driving blood formation. Specifically\, I infer transc ription factor activities in single cells\, revealing how blood stem cells rewire regulatory networks to become distinct mature blood cell types.\n  \n Second\, I present scChIX (single-cell chromatin immunocleavage and u nmixing)\, an integrated machine learning and experimental framework to un derstand the interplay between histone modifications in single cells. scCh IX generates linked maps of active and repressive chromatin\, allowing int egrated analysis of different histone modifications in single cells. These linked maps reveal switching between active and repressive states during development of B cells in the blood.\n  \n Overall\, machine learning met hods integrated into new experimental technologies reveal global gene regu latory principles.\n Join Zoom Meeting https://zoom.us/j/475819702\n Meet ing ID: 475-819-702If you haven't registered for previous QLS webinars\, p lease contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n\ n//indico.ictp.it/event/9653/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9653/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Bidirectional currents in confined driven colloids DTSTART;VALUE=DATE-TIME:20210525T130000Z DTEND;VALUE=DATE-TIME:20210525T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9670@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702\n Meeting ID : 475-819-702\n Abstract:\n It is well known that matter behaves different ly in confined spaces. One example of this is the equilibrium phases of co re softened potential particles\, which can be realized by confining magne tic particles in slit pores thick enough to accommodate one but not two pa rticles.\n I will discuss the rich behavior that appears from experimental ly bringing a similar system out of equilibrium by adding a dynamic magnet ic field. With this strategy\, we can assemble the system in particle pair s\, which can be made to rotate synchronously or asynchronously\, or we ca n obtain buckled phases with soft potentials. By adjusting the packing fra ction\, an intermediate state can be induced in which particle pairs are b roken and reassembled with other particles in a continuous exchange of nei ghbors. By adding a small in-plane component\, this exchange process can b e harnessed to produce a “ceilidh”-like bidirectional current with par ticle transport in a state that is reminiscent of skipping orbits in elect ronic and molecular systems.\n We use Brownian dynamics simulations to sho w that this induced current is due only to the dipolar interaction\, and w e develop a theoretical model that shows two separate transport mechanisms which show excellent agreement with the simulations.If you haven't regist ered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeting\n  \n\n//indico.ictp.it/event/9670/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9670/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Exactly solvable models for high-dimensional machine learning problems DTSTART;VALUE=DATE-TIME:20210608T130000Z DTEND;VALUE=DATE-TIME:20210608T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9678@ictp.it DESCRIPTION:Join Zoom Meeting https://zoom.us/j/475819702\nMeeting ID: 47 5-819-702If you haven't registered for previous QLS webinars\, please cont act qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n \n The pas t decade has witnessed a surge in the development and adoption of machine learning algorithms to solve day-a-day computational tasks. Yet\, a solid theoretical understanding of even the most basic tools used in practice is still lacking. Surprisingly\, many of the “exotic” behaviours of deep neural networks have shown to hold in models as simple as high-dimensiona l linear regression. In this talk\, I will introduce two simple toy models describing learning with correlated features. Next\, I will motivate how these models encompass many learning problems of interest\, e.g. ridge and logistic regression\, kernel methods\, random features\, scattering trans forms and transfer learning. Finally\, I will discuss how these models ca n be used to approximate real data learning curves in some scenarios.\n\n/ /indico.ictp.it/event/9678/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9678/ END:VEVENT BEGIN:VEVENT SUMMARY:Thalamic control of cortical dynamics in a model of flexible motor sequencing DTSTART;VALUE=DATE-TIME:20210622T130000Z DTEND;VALUE=DATE-TIME:20210622T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9680@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702\n Meeting ID : 475-819-702If you haven't registered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n Abstra ct:\n The neural mechanisms that generate an extensible library of motor m otifs and flexibly string them into arbitrary sequences are unclear. We de veloped a model in which inhibitory basal ganglia output neurons project t o thalamic units that are themselves bidirectionally connected to a recurr ent cortical network. We model the basal ganglia inhibitory patterns as si lencing some thalamic neurons while leaving others disinhibited and free t o interact with cortex during specific motifs. We show that a small number of disinhibited thalamic neurons can control cortical dynamics to generat e specific motor output in a noise robust way. Additionally\, a single ‘ preparatory’ thalamocortical network can produce fast cortical dynamics that support rapid transitions between any pair of learned motifs. If the thalamic units associated with each sequence component are segregated\, ma ny motor outputs can be learned without interference and then combined in arbitrary orders for the flexible production of long and complex motor seq uences. Our results suggest that the motor thalamocortical network archite cture is well-suited to implement the stringing of re-usable skilled behav ioral primitives\, and have implications for engineering artificial networ ks for robust hierarchical motor control.\n\n//indico.ictp.it/event/9680/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9680/ END:VEVENT BEGIN:VEVENT SUMMARY:Local entropy\, a selected topic in the physics of deep learning DTSTART;VALUE=DATE-TIME:20210720T130000Z DTEND;VALUE=DATE-TIME:20210720T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9687@ictp.it DESCRIPTION:Join Zoom Meeting https://zoom.us/j/475819702\nMeeting ID: 47 5-819-702If you haven't registered for previous QLS webinars\, please cont act qls@ictp.it by Friday\, 16 July to obtain the PASSWORD for this zoom m eeting.\n\n \n Insight and methods coming from physics have ever since hel ped to improve our understanding and design of neural networks. The interd isciplinary potential of techniques borrowed from statistical physics and information theory\, such as entropic regularizations and renormalization\ , has not yet exhausted its drive in machine learning. After a non-technic al description of such a panorama\, the talk delves into a specific line o f research based on local entropy. Local entropy represents a refinement o f standard entropy and defines a coarse-graining technique helpful in unde rstanding and improving deep neural networks. It allows for further refine ment\, entailing anisotropy in synaptic space and time dependence. As such \, it offers a versatile framework where to study architecture-aware regul arization procedures. Besides\, a scheduling protocol where the regulariza tion is strong in the early-stage of the training and then fades progressi vely away constitutes an alternative to standard initialization procedures . We conclude with some future perspectives\, mainly related to spontaneou sly/explicitly broken scale invariance and hints of spacetime geometry in deep networks.\n\n//indico.ictp.it/event/9687/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9687/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: The universality of epidemics in meta-population scal e DTSTART;VALUE=DATE-TIME:20210907T093000Z DTEND;VALUE=DATE-TIME:20210907T103000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9702@ictp.it DESCRIPTION:Zoom Meeting https://zoom.us/j/475819702\nMeeting ID: 475-819 -702If you haven't registered for previous QLS webinars\, please contact q ls@ictp.it to obtain the PASSWORD for this zoom meeting.\n We introduce a generic meta-population framework by which the spreading dynamic in both l evels of inter-population and intra-populations evolve. Then we show how t o estimate A) where\, B) when the dynamic had started and C) calculate th e cut-off errors in this frame. Finally we redefine the effective distance concept and show that the effective distance vs overtaking time\, when th e inter-population dynamic overtakes the intra-population dynamic\, prese nts a universal geometry\, namely a line\, for those neighbours of the ori gin whose reachability via intermediate nodes are negligible in compariso n to the direct effective paths. This line has a universal slope for any disease or network of mobility or origin node. The empirical data of COVI D-19 epidemics in Iran and the United States of America as well as the H1N 1 pandemic in the world confirms our theory.\n\n\n//indico.ictp.it/event/9 702/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9702/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Thermodynamic Machine Learning through Maximum Work Production DTSTART;VALUE=DATE-TIME:20211019T120000Z DTEND;VALUE=DATE-TIME:20211019T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9719@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702\n Meeting ID : 475-819-702\nIf you haven't registered for previous QLS webinars\, pleas e contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n Adap tive systems—such as a biological organism gaining survival advantage\, an autonomous robot executing a functional task\, or a motor protein trans porting intracellular nutrients—must model the regularities and stochast icity in their environments to take full advantage of thermodynamic resour ces. Analogously\, but in a purely computational realm\, machine learning algorithms estimate models to capture predictable structure and identify i rrelevant noise in training data. This happens through optimization of per formance metrics\, such as model likelihood. If physically implemented\, i s there a sense in which computational models estimated through machine le arning are physically preferred? We introduce the thermodynamic principle that work production is the most relevant performance metric for an adapti ve physical agent and compare the results to the maximum-likelihood princi ple that guides machine learning. Within the class of physical agents that most efficiently harvest energy from their environment\, we demonstrate t hat an efficient agent’s model explicitly determines its architecture an d how much useful work it harvests from the environment. We then show that selecting the maximum-work agent for given environmental data corresponds to finding the maximum-likelihood model. This establishes an equivalence between nonequilibrium thermodynamics and dynamic learning. In this way\, work maximization emerges as an organizing principle that underlies learni ng in adaptive thermodynamic systems.\n\n//indico.ictp.it/event/9719/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9719/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Computational Methods in the Discovery of Bioactive N atural Products DTSTART;VALUE=DATE-TIME:20211026T130000Z DTEND;VALUE=DATE-TIME:20211026T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9737@ictp.it DESCRIPTION:\n In-person meeting: Common area\, exSISSA bldg\, 2nd floor\n and\n  Zoom Meeting https://zoom.us/j/475819702\n Meeting ID: 475-819-7 02\n  If you haven't registered for previous QLS zoom webinars\, please c ontact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n  \n\n Traditionally\, drug design and development is quite lengthy\, costly\, an d a challenging process. On average\, it takes about 12–15 years\, costi ng $2.6 billion for a new drug to the approval stage [1]. The approval rat e of new drugs to enter clinical market is only 12% [1]\; unfortunately\, some drugs are banned from clinical uses due to undesirable pharmacokineti c profiles. On the other hand\,  natural products are safe and have serve d human life as medications for centuries. In Tanzania\, a number of natur ally derived compounds and extracts have been tested or used as potential remedies against many diseases including COVID-19. Understanding their mod e of action is an important step in drug discovery and development. Recent advances in computational power and methods have accelerated the process of drug design\, discovery\, and development\, by studying mode of action\ , pharmacological\, biological\, and pharmacokinetic properties of lead mo lecules before entering pre-clinical or clinical trials [2-4]. Computation al approaches help in identifying molecules demonstrating both drug-like a nd undesirable properties at an early stage\, hence reducing both time and cost related to drug development. In this talk\, the application of diffe rent computational methods applied in the discovery of bioactive natural p roducts will be discussed. First\, the talk will discuss various classes o f natural products and their biological activities. Then\, different compu tational methods used in the study of natural products will be discussed. Finally\, the discovery of potent anti-SARS-CoV-2 natural products current ly in clinical trial will be highlighted\, among others.Reference\n [1] Su llivan T (2019) A tough road: cost to develop one new drug is $2.6 billion \; approval rate for drugs entering clinical development is less than 12%. Policy & Medicine.\n [2] De Vivo M\, Masetti M\, Bottegoni G\, Cavalli A (2016) Role of molecular dynamics and related methods in drug discovery. J Med Chem 59(9):4035–4061.\n [3] Amaro RE\, Schnaufer A\, Interthal H\, Hol W\, Stuart KD\, McCammon JA (2008) Discovery of drug-like inhibitors o f an essential RNA-editing ligase in Trypanosoma brucei. Proc Natl Acad Sc i 105(45):17278–17283.\n [4] Brooijmans N (2009) Docking methods\, ligan d design\, and validating data sets in the structural genomic era. Structu ral Bioinformatics. Wiley\, New York\, NY\, USA\, pp 635–663.*Please not e: in-person attendance is allowed to visitors who have a EU Digital COVID Certificate\, or are in possession of a negative rapid antigen test (with in the previous 48 hours) or PCR test (within the previous 72 hours). \n\n //indico.ictp.it/event/9737/ LOCATION:Zoom Meeting and Common area\, exSISSA bldg\, 2nd floor URL://indico.ictp.it/event/9737/ END:VEVENT BEGIN:VEVENT SUMMARY:Joint ICTP-SISSA seminar - Tensor decomposition of higher-order co rrelations by nonlinear Hebbian plasticity DTSTART;VALUE=DATE-TIME:20211130T150000Z DTEND;VALUE=DATE-TIME:20211130T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9747@ictp.it DESCRIPTION:\n Join Zoom Meeting https://sissa-it.zoom.us/j/6902850337\n\n \n This seminar will be held jointly with Theoretical and Scientific Data Science (TSDS) at SISSA\n\n \n Biological synaptic plasticity exhibits no nlinearities that are not accounted for by classic Hebbian learning rules. Here\, we introduce a simple family of generalized nonlinear Hebbian lear ning rules. We study the computations implemented by their dynamics in the simple setting of a neuron receiving feedforward inputs. These nonlinear Hebbian rules allow a neuron to learn tensor decompositions of its higher- order input correlations. The particular input correlation decomposed and the form of the decomposition depend on the location of nonlinearities in the plasticity rule. For simple\, biologically motivated parameters\, the neuron learns eigenvectors of higher-order input correlation tensors. We prove that tensor eigenvectors are attractors and determine their basins o f attraction. We calculate the volume of those basins\, showing that the d ominant eigenvector has the largest basin of attraction. We then study arb itrary learning rules and find that any learning rule that admits a finite Taylor expansion into the neural input and output also has stable equilib ria at generalized eigenvectors of higher-order input correlation tensors. Nonlinearities in synaptic plasticity thus allow a neuron to encode highe r-order input correlations in a simple fashion.\n\n\n//indico.ictp.it/even t/9747/ LOCATION: URL://indico.ictp.it/event/9747/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Data-driven Modeling of Animal Behavior DTSTART;VALUE=DATE-TIME:20220201T140000Z DTEND;VALUE=DATE-TIME:20220201T150000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9855@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702\n Meeting ID : 475-819-702\nIf you haven't registered for previous QLS webinars\, pleas e contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n In t his talk\, I will describe recent work on analyzing behavioral dynamics us ing the theory of nonlinear dynamics. I will first describe ways to conver t time-series observations into a phase-space\, which is a maximally predi ctive geometric representation of the data. The phase-space\, once properl y reconstructed\, contains all dynamically relevant information available in the measurements. Next\, I will describe methods to extract meaningful information from the phase space by analyzing its local geometry and topol ogy. Applications include separating continuous behaviors into discrete mo tifs\, identifying salient events\, extracting long time-scale structure\, quantifying behavioral variability and studying coupling between sub-syst ems.\n\n  \n\n  \n\n References:\n\n  \n\n 1. Ahamed\, Tosif\, Antonio C. Costa\, and Greg J. Stephens. "Capturing the continuous complexity of b ehaviour in Caenorhabditis elegans." Nature Physics 17.2 (2021): 275-283.\ n\n 2. Costa\, Antonio Carlos\, et al. "Maximally predictive ensemble dyna mics from data." arXiv preprint arXiv:2105.12811 (2021).\n\n  \n\n//indic o.ictp.it/event/9855/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9855/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS SEMINAR - Activity driven pattern formation in microtubule net work and cell membrane DTSTART;VALUE=DATE-TIME:20220202T130000Z DTEND;VALUE=DATE-TIME:20220202T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9858@ictp.it DESCRIPTION:\n Self assembly of different globular and filamentous protein s in the eukaryotic cell throughout the entire cell cycle is ubiquitous an d helps a cell to perform various tasks such as migration from one place t o another\, intake of food from the outside world\, cell division\, etc. T he organization of individual microtubule filaments into bundles during th e formation of the mitotic spindle is an example of such a phenomenon when a cell enters mitosis. At the beginning of prometaphase\, experimentally it is found that at the midplane of a vertically oriented spindle\, microt ubules form a mist-like distribution which transits to well-formed droplet like structures as time progresses. We construct a free energy descriptio n of this system using the density of microtubules and cross-linking prote ins as field variables considering attractive and repulsive interactions b etween them. Further\, dynamical equations governing the time evolution of density fields are obtained by minimizing the free energy and incorporati ng a non-equilibrium process of microtubule polymerization and depolymeriz ation. Linear stability analysis shows that the system transits from homog eneous distribution of microtubules to a phase consisting of multiple bund les of microtubules once the density of cross-linking proteins crosses a t hreshold and the bundles do not collapse into a single large bundle over t ime.\n  On the other hand\, pattern formation on the cell membrane is lin ked with the contractile active force generated by the actomyosin cortex b eneath it and outward push coming due to polymerization of filamentous act in attached to the membrane surface. Membrane associated proteins act as t he nucleating centers of actin polymerization\, and the proteins couple to the membrane surface in a curvature dependent manner modifying the local bending modulus of the membrane. These competing equilibrium and non-equil ibrium forces acting on cell membrane allow a cell to deform\, and mediate cell motility and division. We use coupled evolution of fields to perform linear stability analysis and numerical calculations. As activity overcom es the stabilizing factors such as surface tension and bending rigidity\, the spherical membrane shows instability towards pattern formation\, local ized pulsation\, and running pulsation between poles.\n\n//indico.ictp.it/ event/9858/ LOCATION:ICTP\, Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9858/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS-CMSP Joint Seminar - Homogeneization of Markovian process and application to quantum continuous measurement DTSTART;VALUE=DATE-TIME:20220222T143000Z DTEND;VALUE=DATE-TIME:20220222T153000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9880@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702\n\n If you h aven't registered for previous QLS webinars\, please contact qls@ictp.it t o obtain the PASSWORD for this zoom meeting.\n In many physical systems\, we are faced with an interaction between mechanisms that evolve at very di fferent time scales. In this seminar\, we will consider Markovian processe s with such characteristics\, and we will exploit the time scale separatio n to obtain an effective Markovian dynamics on a reduced state space. In p articular\, compared to existing studies\, we will introduce a more genera l formalism.\n\n Finally\, as a privileged example\, we consider the limit of the strong continuous measure of quantum systems described by a Belavk in diffusive equation.\n\n \n  \n\n//indico.ictp.it/event/9880/ LOCATION:Zoom Meeting and ICTP Central Area\, 2nd floor\, old SISSA buildi ng URL://indico.ictp.it/event/9880/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Sciences involving humans DTSTART;VALUE=DATE-TIME:20220303T133000Z DTEND;VALUE=DATE-TIME:20220303T143000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9903@ictp.it DESCRIPTION:\n This informal talk will discuss quantitative approaches to problems involving human dynamics\, such as 1) fitting Langevin dynamics t o time-series of prehistorical societies\; 2) Predicting the next pandemic using war games with human role players\; and (3) Block chain\, ecosystem s of AIs\, and the arrival of the singularity.\n\n//indico.ictp.it/event/9 903/ LOCATION:ICTP Common area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9903/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Binary sequences and individual-based models applied to evolution DTSTART;VALUE=DATE-TIME:20220330T123000Z DTEND;VALUE=DATE-TIME:20220330T133000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9920@ictp.it DESCRIPTION:\n Individual-based models (IBMs)\, unlike the mean-field ones \, are commonly not described by equations\, but instead by well-defined r ules\, which are meant to be ran in a computer. Although widely used in ma ny biological fields\, it is not easy to achieve general results\, once th ey can bear many tunable parameters whose interplay can be very complex. I n this talk\, I aim to introduce three interesting models in which a mathe matical description can be formulated and some analytical results can be f ound.\n Considering binary strings as a proxy for individuals’ genetic s equence\, it is possible to analyze their evolution under different scenar ios. The so-called Derrida-Higgs model\, which is an IBM of this type\, mi mics the species formation process in a sympatric community\; from this\, a model of coevolution between mitochondrial and nuclear DNA has been prop osed by Príncepe and de Aguiar (2020)\, which will also be covered during the talk.\n Moreover\, binary strings have also been used to model RNA vi ruses going on epidemic\, and together with network theory they can provid e a powerful framework for the study of viral evolution.\n  \n\n//indico. ictp.it/event/9920/ LOCATION:ICTP Central Area\, 2nd floor\, ex SISSA building URL://indico.ictp.it/event/9920/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Phase separation in active matter systems DTSTART;VALUE=DATE-TIME:20220405T130000Z DTEND;VALUE=DATE-TIME:20220405T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9914@ictp.it DESCRIPTION:\n The field of active matter studies materials composed of se lf-propelled entities\, which can be either biological or synthetic. Examp les of such materials stretch across length-scales\, from nanorobots to hu mans\, and provide a unique plethora of collective behaviors that are inac cessible in equilibrium materials\, ranging from highly ordered bird flock s to phase separation in absence of any attraction among particles. The ro le of theoretical research in active matter is thus not only to explain na tural phenomena but also to deliver robust models that allow to unveil whi ch novel phenomenology we should expect. \n  \n The talk will start wit h a brief overview of active matter and of my recent contributions to it.  I will then focus on phase separation in active systems. This is one of the most fundamental collective phenomena arising at high density\, and it takes place in systems as diverse as bacterial suspensions and biological tissues. I will show that much understanding can be drawn by generalising the classical field theoretical description of liquid-liquid phase separa tion (Model B\, or phi^4 theory) to an active context. After discussing th e thermodynamics of such model\, I will show that activity induces phase s eparated phases that have qualitatively new features with respect to thos e encountered in passive systems. This surprising fact can be rationalised by generalising the concept of interfacial tension\, crucial in passive p hase separation\, to an active context. Next\, we will ask how to control these new forms of phase separation in terms of microscopically tunable pa rameters\; although this question is still largely open\, I will describe the recent progresses we made by studying minimal models of active particl es.\n  \n\n//indico.ictp.it/event/9914/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9914/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - An intrinsic dimension estimator for discrete-metric spaces: formalization and first applications DTSTART;VALUE=DATE-TIME:20220406T130000Z DTEND;VALUE=DATE-TIME:20220406T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9921@ictp.it DESCRIPTION:\n Real world datasets characterised by discrete features are ubiquitous: From categorical surveys to clinical questionnaires\, from unw eighted networks to genomic strands. Nonetheless\, the development of meth ods to treat data with discrete features lags behind\, particularly concer ning geometric and manifold learning approaches. Due to the lack of such t ools\, the analysis of aforementioned dataset still relies on algorithms d eveloped for continuous spaces\, inevitably introducing approximations\, e rror and biases. In this work\, starting from the appropriate definition o f volumes on lattices\, we develop a very simple\, yet effective\, routine to estimate the intrinsic dimension of datasets naturally described by di screte metric spaces. Besides\, our id estimator allows to explicitly sele ct the scale at which the id is computed\, an important property that is h ardly provided even in estimators for continuous spaces. We assess the val idity of the new estimator on artificial datasets against a state of the a rt continuous estimator and then apply it to a controlled-id spin system a s well as to an ensemble of genomic sequences\n\n//indico.ictp.it/event/99 21/ LOCATION:ICTP Central Area\, 2nd floor\, ex SISSA building URL://indico.ictp.it/event/9921/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS webinar - Optimal Mass Transport meets Stochastic Thermodynami cs: a geometric approach to thermodynamic cycles with anisotropic fluctuat ions DTSTART;VALUE=DATE-TIME:20220426T140000Z DTEND;VALUE=DATE-TIME:20220426T150000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9934@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702\n Meeting ID : 475-819-702If you haven't registered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeting.\n \n \ n\n In contrast to the classical concept of a Carnot engine that alternate s contact between heat baths of different temperatures\, naturally occurri ng processes usually harvest energy from anisotropy\, being exposed simult aneously to chemical and thermal fluctuations of different intensities. In these cases\, the enabling mechanism responsible for transduction of ener gy is the presence of a non-equilibrium steady state (NESS). A suitable st ochastic model for such a phenomenon is the Brownian gyrator -- a two-degr ee of freedom stochastically driven system that exchanges energy and heat with the environment. In this context\, we exploit the relationship betwee n optimal mass transport and stochastic thermodynamics to present a geomet ric view of the energy harvesting mechanism\, that entails a forced period ic trajectory of the system state on the thermodynamic manifold. Dissipati on and work output are expressed accordingly as path integrals of a contro lled process\, and fundamental limitations on power and efficiency are exp ressed in geometric terms via a relationship to an isoperimetric problem. The analysis presented provides guiding principles for building autonomous engines that extract work from anistropic fluctuations.\n  \nReferences: \nO. Movilla Miangolarra\, A.Taghvaei\, R. Fu\, Y.Chen\, T. T. Georgiou\, Energy harvesting from anisotropic fluctuations\, Phys. Rev. E (2021)\nO. Movilla Miangolarra\, A.Taghvaei\, Y.Chen\, T. T. Georgiou\, Geometry of f inite-time thermodynamic cycles with anisotropic thermal fluctuations\, ar xiv:2203.12483 (2022)\nO. Movilla Miangolarra\, A.Taghvaei\, Y.Chen\, T. T . Georgiou\, Thermodynamic engine powered by anisotropic fluctuations\, Ar xiv:2203.07573 (2022)\n \n\n//indico.ictp.it/event/9934/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9934/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS webinar - Uncertainty relation for enzymes\, resetting systems \, and first passage time processes with irreversible transitions DTSTART;VALUE=DATE-TIME:20220503T120000Z DTEND;VALUE=DATE-TIME:20220503T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9941@ictp.it DESCRIPTION:\n Join Zoom Meeting https://zoom.us/j/475819702\n Meeting ID : 475-819-702If you haven't registered for previous QLS webinars\, please contact qls@ictp.it to obtain the PASSWORD for this zoom meeting\n\n \n Th e thermodynamic uncertainty relation (TUR) shows that the fluctuations of currents are bounded from below by the inverse entropy production. When on e or more transitions are completely irreversible the entropy production d iverges\, and the TUR becomes non-informative. We show that additional bou nds\, which mix entropic and dynamic contributions\, hold for such process es. Crucially\, these new bounds are tighter in the presence of irreversib le transitions. A steady-state process with resetting\, enzymes and a tran sient first-passage problem are used as examples. We also discuss the conn ections between the bounds and the Aldous-Shepp bound that is often used i n statistical kinetics. \n\n//indico.ictp.it/event/9941/ LOCATION:Zoom Meeting URL://indico.ictp.it/event/9941/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Monthly Colloquium Series - Computational Physics of Active Fi laments\, Membranes\, and Cells DTSTART;VALUE=DATE-TIME:20220511T090000Z DTEND;VALUE=DATE-TIME:20220511T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9951@ictp.it DESCRIPTION:\n Zoom registration: https://bit.ly/QLS11May\n Active matter exhibits a wealth of emergent non-equilibrium behaviors. A paradigmatic ex ample is the interior of biological cells. The intimate coupling of active forces\, thermal noise\, hydrodynamic interactions\, and polymeric connec tivity implies the emergence of novel structural and dynamical features.\n Different propulsion mechanisms capture the physics of active polymeric s ystems\, such as chains of active Brownian particles\, and polar tangentia lly propelled laments. This leads to interesting single-particle behavior\ , such as a softening of active Brownian laments at intermediate levels of activity. At high polymer densities\, collective dynamics is characterize d by active turbulence.\n Vesicles with internal active components are hig hly simplifed models of cells. Here\, the active components lead to enhanc ed fluctuations and an intimate coupling of propulsion forces\, membrane d eformability\, cell shape\, and cell sensing and reactivity.\n In all thes e systems\, computational models of active matter play an essential role t o elucidate their non-equilibrium behavior.\n  \n\n//indico.ictp.it/event /9951/ LOCATION:Zoom URL://indico.ictp.it/event/9951/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS-CMSP Joint Seminar - Thermal relaxation of a hot nano/bio-part icle into the water from an atomistic classical view DTSTART;VALUE=DATE-TIME:20220613T080000Z DTEND;VALUE=DATE-TIME:20220613T090000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9976@ictp.it DESCRIPTION:\n Knowing the thermal relaxation of amino acids as the buildi ng blocks of bio-molecules and proteins can provide a better understanding of how a bio-molecule dissipates heat when chemical and photochemical rea ctions occur at specific spots. In this talk\, I will show our recent resu lts about the thermal relaxation of 20 naturally occurring amino acids in water and in lysozyme protein using transient nonequilibrium molecular dyn amics [1]. The relationship between the cooling time of amino acids to the solvent-accessible surface area of a sample protein in water is discussed . Our finding might help to analyze the function of water-surrounded amino acids\, proteins\, and bio-molecules in the face of temperature gradients \, which have a wide range of applications\, including cancer thermotherap y. I also present how to calculate the interfacial thermal conductance (Ka pitza conductance) between different nanoparticles and water by equilibriu m and non-equilibrium simulation techniques in both steady and unsteady st ate conditions [2]. These nanoparticles include gold\, silver\, graphene f lakes and titanium dioxide\, and amino acids. The effect of size\, hydroph ilicity\, hydrophobicity\, the contribution of van der Waals and polar int eractions\, as well as the effects of electric charges on nanoparticles an d the presence of ions in water are investigated [3-5].\n\n [1] H Hamzi\, A Rajabpour\, É Roldán\, A Hassanali\, “Learning the Hydrophobic\, Hyd rophilic\, and Aromatic Character of Amino Acids from Thermal Relaxation a nd Interfacial Thermal Conductance” J. Phys. Chem. B. 126 (2022) 670-678 \n  \n [2] A Rajabpour\, R Seif\, S Arabha\, MM Heyhat\, S Merabia\, A Ha ssanali\, “Thermal transport at a nanoparticle-water interface: A molecu lar dynamics and continuum modeling study” J. of Chem. Phys. 150 (2019) 114701   \n  \n [3] R Rabani\, MH Saidi\, L Joly\, S Merabia\, A Rajab pour\, “Enhanced local viscosity around colloidal nanoparticles probed b y equilibrium molecular dynamics simulations” J. of Chem. Phys. 155 (202 1) 174701\n  \n [4] F Jabbari\, A Rajabpour\, S Saedodin\, S Wongwises\, “Effect of water/carbon interaction strength on interfacial thermal resi stance and the surrounding molecular nanolayer of CNT and graphene flake ” J. of Molecular Liquids 282 (2019) 197-204.\n            \n [5] MM Heyhat\, M Abbasi\, A Rajabpour\, “Importance of nanolayer forma tion in nanofluid properties: Equilibrium molecular dynamic simulations fo r Ag-water nanofluid” J. Molecular Liquids 333 (2021) 115966\n\n//indico .ictp.it/event/9976/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9976/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS-CMSP Joint Seminar - A coarse-grained approach to get viruses into your desktop computer: A close look into Zika virus structure and dyn amics DTSTART;VALUE=DATE-TIME:20220613T090000Z DTEND;VALUE=DATE-TIME:20220613T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9979@ictp.it DESCRIPTION:\n Physical Virology is a relatively new discipline aimed to u nravel the physical forces that rule the assembly and destabilization of v iral particles. Computational approaches can provide structural and dynami cal insights into those processes. However\, the computational cost of sim ulating entire viral particles constituted by millions of atoms is prohibi tive without the use of supercomputer resources. We have developed a plug& play coarse-grained force field for biomolecules including topologies and interaction parameters for proteins\, lipids\, nucleic acids and aqueous s olvent named SIRAH (Southamerican Initiative for a Rapid and Accurate Hami ltonian). Our force field grants over 2 orders of magnitude speed up\, mak ing the simulation of viruses affordable to desktop computers.\n We used t his force field to characterize the dynamics of Zika virus\, revealing unf oreseen protein-lipid interactions that contribute to store significant te nsile energy in the viral membrane\, proposing it as an important factor i n the spring-loaded configuration of the mature virion.\n  References\n [ 1] Soñora M\, Martínez L\, Pantano S\, Machado MR\, Wrapping Up Viruses at Multiscale Resolution: Optimizing PACKMOL and SIRAH Execution for Simul ating the Zika Virus. JCIM\, 2021\, 61:408\n [2] Machado M\, Gonzalez HC\, Pantano S. MD Simulations of Virus-Like Particles with Supra CG solvation affordable to desktop computers. JCTC. 2017\, 13: 5106.\n [3] Soñora M\, Barrera EE\, Zonta F\, Pantano S\, The stressed life of a lipid in the Zi ka Virus membrane\, BBA – Membranes\, 2022\, 1864 (1)\, 183804.\n\n//ind ico.ictp.it/event/9979/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9979/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Network Resonance for Selective Neural Entrainment DTSTART;VALUE=DATE-TIME:20220628T120000Z DTEND;VALUE=DATE-TIME:20220628T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9994@ictp.it DESCRIPTION:\n We model the brain as a network of Kuramoto loosely coupled oscillators\, where node frequencies are determined by physiological obse rvations\,  and the edges given by the human connectome. Global synchrony can be calculated from Kuramoto’s Order parameter\, essentially the ave rage phase over all the nodes.  We argue that structural and functional s ubnetworks with the global network possess frequency characteristics akin to network resonance in electric circuits\, such that selective entrainmen t of networks could be achieved by externally driving the network close it s ‘resonance’ frequency – and that such entrainment would have behav ioural consequences on the tasks associated with the selected network.\n T his framework could be used to selectively facilitate cognitive processing by enhancing neural oscillations affected\, for example\, by neurodegener ation.Zoom Meeting ID to attend the seminar online: 475-819-702\n Join Zoo m Meeting:https://zoom.us/j/475819702\n  \n\n If you haven't registered f or previous QLS webinars\, please contact qls@ictp.it to obtain the passwo rd for this zoom meeting.\n\n \n  \n\n//indico.ictp.it/event/9994/ LOCATION:ICTP Central area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9994/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Novel mechanism for oscillations in catchbonded motor -filament complexes DTSTART;VALUE=DATE-TIME:20220705T120000Z DTEND;VALUE=DATE-TIME:20220705T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9993@ictp.it DESCRIPTION:\n Catchbonding\, characterized by the increase of the molecul ar bond lifetime in response to applied forces is observed for biological macromolecules\, including motor proteins within cells. Hitherto\, the pri mary biological functional role of catchbonding was thought to be that of improving adhesion properties of cells to surfaces when subject to tensile forces. However\, our study reveals that allosteric deformations in dynei n motors\, which lead to the catchbonding\, can also manifest as a generic mechanism to generate spontaneous mechanical oscillations in motor-cytosk eletal filament complexes. The resultant phase diagram of such motor-filam ent systems\, characterized by force-induced allosteric deformations exhib its bistability and sustained limit-cycle oscillations in biologically rel evant regimes\, such as for catchbonded dynein. The results elucidate the central role of this mechanism in fashioning a distinctive stability behav ior and oscillations in motor-filament complexes such as mitotic spindles. In essence\, this study uncovers a crucial functional role of the biologi cal catchbonding in intracellular processes ranging from spindle oscillati ons during cell division to intracellular transport.Reference:\n Biophys. J\, 120\, 4129 (2021)\n  \n\n//indico.ictp.it/event/9993/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/9993/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Meeting - The adjacent possible of stochastic thermodynamics DTSTART;VALUE=DATE-TIME:20220711T120000Z DTEND;VALUE=DATE-TIME:20220713T160000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-9991@ictp.it DESCRIPTION:Speakers:\n\n Edgar Roldan (ICTP)\n \n Matteo Marsili (ICTP) \n \n Gonzalo Manzano Paule (IFISC\, Mallorca)\n \n Jan Korbel (Medical University\, Vienna)\n \n Andres Ortiz (Santa Fe Institute)\n \n Gülce Kardes (ICTP)\n \n David Wolpert (Santa Fe Institute)\n \n Farita Tasnim (MIT)\n \n Sebastian Goldt (SISSA)\n \n Jean Barbier (ICTP)\n \n Andre a Solfanelli (SISSA)\nTopics:\n\n Stochastic thermodynamics for computer science and Shannon coding\n \n Statistical physics in new domains\n \n Learning\, neurobiological systems\n\n \n \n \n \n Join Zoom Mee ting https://zoom.us/j/475819702 Meeting ID: 475-819-702 \n \n If y ou haven't registered for previous QLS webinars\, please contact qls@ictp. it to obtain the PASSWORD for this zoom meeting. \n \n \n \n \n  \n\n/ /indico.ictp.it/event/9991/ LOCATION:ICTP Leonardo Building - Lecture Room B URL://indico.ictp.it/event/9991/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Metabolic interactions in natural and synthetic micr obial communities DTSTART;VALUE=DATE-TIME:20220718T090000Z DTEND;VALUE=DATE-TIME:20220718T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10011@ictp.it DESCRIPTION:\n Metabolism operates as a multiscale process that induces in teractions between microbial species and leads to division of labor and cr oss-feeding in complex microbial communities. By combining dynamic flux ba lance modeling with diffusion equations\, we model the spatio-temporal dyn amics of communities in complex environments\, starting from the genomes o f individual species. We envisage that these approaches will be increasing ly valuable for understanding natural communities (e.g.\, ocean and soil c onsortia)\, and for guiding targeted synthetic ecology experiments with ar tificial communities that can be used both as benchmark for our models and as avenues for environmental remediation and metabolic engineering.\n\n// indico.ictp.it/event/10011/ LOCATION:ICTP Central Area\, Second floor\, old SISSA Building URL://indico.ictp.it/event/10011/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Two notes on the foundations of statistical mechanic s: objectivity and the origin of giant fluctuations DTSTART;VALUE=DATE-TIME:20220719T120000Z DTEND;VALUE=DATE-TIME:20220719T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10009@ictp.it DESCRIPTION:\n Boltzmann’s explanation of irreversibility is based on th e concept of macro-states and the definition of entropy as the logarithm o f the volume in phase space of the region of micro-states compatible with a given macro-state. The explanation\, however\, lacks an objective (i.e. non arbitrary) definition of macro-states and of the crossover between mic ro- and macro-scales. Here we show that this problem can be solved by refo rmulating Boltzmann’s explanation in terms of observables relaxing from giant fluctuations. We show that the irreversible behavior of an observabl e is a fully objective property and has nothing to do with its micro- or m acroscopic nature. In fact\, we will show a situation where a system exhib its irreversibility at the micro-scale and reversibility at the macro-scal e. In the second part of the talk\, we propose a mechanism for creating gi ant fluctuations of an observable (hence\, irreversibility) based on metas table states induced by symmetry breaking.\n\n//indico.ictp.it/event/10009 / LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10009/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Diffusive growth sourced by topological defects DTSTART;VALUE=DATE-TIME:20220721T090000Z DTEND;VALUE=DATE-TIME:20220721T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10008@ictp.it DESCRIPTION:\n In this talk\, we develop a minimal model of morphogenesis of a surface where the dynamics of the intrinsic geometry is diffusive gro wth sourced by topological defects. We show that a positive (negative) def ect can dynamically generate a cone (hyperbolic cone). We analytically exp lain features of the growth profile as a function of position and time\, a nd predict that in the presence of a positive defect\, a bump forms with h eight profile h(t) ~ t^(1/2) for early times t. To incorporate the effect of the mean curvature\, we exploit the fact that for axisymmetric surfaces \, the extrinsic geometry can be deduced entirely by the intrinsic geometr y. We find that the resulting stationary geometry\, for polar order and sm all bending modulus\, is a deformed football.\n\n We apply our framework t o various biological systems. In an ex-vivo setting of cultured murine neu ral progenitor cells\, we show that our framework is consistent with the o bserved cell accumulation at positive defects and depletion at negative de fects. In an in-vivo setting\, we show that the defect configuration consi sting of a bound +1 defect state\, which is stabilized by activity\, surro unded by two -1/2 defects can create a stationary ring configuration of te ntacles\, consistent with observations of a basal marine invertebrate Hydr a.\n\n//indico.ictp.it/event/10008/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10008/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - The impact of microbial diseases of corals under mac roalgal toxicity and overfishing DTSTART;VALUE=DATE-TIME:20220803T120000Z DTEND;VALUE=DATE-TIME:20220803T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10022@ictp.it DESCRIPTION:\n \n \n \n Competition between macroalgae and corals fo r occupying the available space in sea bed is an important ecological proc ess underlying coral-reef dynamics. We investigate coral- macroalgal phase shift in presence of macroalgal allelopathy and microbial infection on co rals by means of an eco-epidemiological model under the assumption that th e transmission of infection occurs through both contagious and non-contagi ous pathways. We found that the system is capable of exhibiting the existe nce of two stable configurations by saddle-node bifurcations. \n \n \n\n\ n  \n\n//indico.ictp.it/event/10022/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10022/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Evolution in spatially heterogeneous environments DTSTART;VALUE=DATE-TIME:20220804T090000Z DTEND;VALUE=DATE-TIME:20220804T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10023@ictp.it DESCRIPTION:\n Generally\, resources and hazards are not distributed unifo rmly. Thus\, environmental interactions are a major factor in determining the success of a new mutant in structured populations. Spatial variations of concentration of resources change the fitness of competing strategies locally and thus can drastically change the outcome of evolutionary proces s in unintuitive ways. I discuss fixation probability and fixation time f or a Moran birth-death process as the fitness heterogeneity and its patte rn vary.\n  \n\n//indico.ictp.it/event/10023/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10023/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Passive and active cluster crystals DTSTART;VALUE=DATE-TIME:20220824T090000Z DTEND;VALUE=DATE-TIME:20220824T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10031@ictp.it DESCRIPTION:\n We study a system of  particles with soft repulsive intera ctions that lead to a cluster-crystal phase\n (solid in which the unit cel l is occupied by a cluster of particles)  in two dimensions. We analyze b oth situations: when particles are passive and active. For the case of act ive particles we use two different modelizations of the active force: Acti ve Brownian particles\, and Ornstein-Uhlenbeck particles\, and focus on an alogies and differences between them.\n\n//indico.ictp.it/event/10031/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10031/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Physics-based Simulations in Computer-Aided Drug Dis covery against neglected diseases DTSTART;VALUE=DATE-TIME:20220825T120000Z DTEND;VALUE=DATE-TIME:20220825T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10032@ictp.it DESCRIPTION:\n This presentation outlines a low cost use of Molecular Mech anics-Poisson Boltzmann approach (MM-PB) to describe interactions between inhibitors and their macromolecular targets taking account of conformation al adaptation upon ligand binding. Computed Gibbs Free energy (Protein-Inh ibitor) of complex formation deltaG (enthalpic\, entropic and solvation co mponents) for a set of inhibitors (training set) explains their biological activity change and identifies their bioactive bound conformation. The su bsequent Pharmacophore (built from the bioactive bound conformation) serve s to explore the virtual chemical subspace in the search of new more poten t analogues of studied inhibitors. Five case studies related to neglected diseases will be presented.\n\n//indico.ictp.it/event/10032/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10032/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS-HECAP Joint seminar: Statistical mechanical perspectives on so me cosmological problems DTSTART;VALUE=DATE-TIME:20220907T120000Z DTEND;VALUE=DATE-TIME:20220907T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10040@ictp.it DESCRIPTION:\n We review some well-known paradoxes in cosmology and give t hem a statistical mechanics reading.  Puzzles to be touched include the h orizon and the flatness problem\, the information paradox\, the dark energ y problem and the possible origin of the so-called space roar. Each time\, we emphasize the role of statistical arguments to complement the dynamica l understanding.  Exciting directions for nonequilibrium (statistical) co smology are briefly added\, especially concerning geometric dissipation.\n \n//indico.ictp.it/event/10040/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/10040/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Structure and dynamics of domains in lipid monolayer s DTSTART;VALUE=DATE-TIME:20220915T120000Z DTEND;VALUE=DATE-TIME:20220915T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10045@ictp.it DESCRIPTION:\n A challenging problem in the biophysics of cell membranes i s the characterization of the lateral organization of their domains. These are made up of lipids in an ordered phase and are dispersed in a more dis ordered continuous phase. A widely used experimental model to study these systems are the lipid monolayers at the water-air interface. Based on thes e experiments and using Brownian simulations\, in this talk we will see ho w the dipolar interaction between domains and the polydispersity in sizes affect the structure of monolayers with phase coexistence and the dynamics of their domains. We will also present a new way to estimate the dipole r epulsion directly from the experimental results.\n\n//indico.ictp.it/event /10045/ LOCATION:ICTP Common area\, second floor\, former SISSA building URL://indico.ictp.it/event/10045/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Mpemba effect in non-equilibrium systems DTSTART;VALUE=DATE-TIME:20220916T090000Z DTEND;VALUE=DATE-TIME:20220916T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10039@ictp.it DESCRIPTION:\n Mpemba effect is a counter-intuitive relaxation phenomenon where an initially prepared hotter system equilibrates faster than the one prepared at a lower temperature. Although common sense suggests that the colder one should equilibrate faster\, this is an apparent paradox that go es back to the time of Aristotle\, Rene Descartes\, and Sir Francis Bacon. The effect was first discovered in the case of water\, where it is descri bed in terms of temperature as a measure. However\, it turns out the effec t is more general and can be studied in the context of relaxation dynamics in non-equilibrium systems. In this talk\, we describe the Mpemba effect in two such systems namely driven granular gases and a colloidal Langevin system. The two examples provide a strong interplay between theory and exp eriment and can render a better understanding of the existence of the Mpem ba effect. An exact analysis determining the criteria for the existence of the Mpemba effect in both systems will be presented. We will underpin the role of anisotropy in the kinetic energies and the presence of metastable states in the energy landscape that cause this counter-intuitive relaxati on.\n\n//indico.ictp.it/event/10039/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10039/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Solubilisation of membrane proteins and lipids into functional bilayer nanodiscs by poly(diisobutylene-alt-maleic acid) and ot her polymers DTSTART;VALUE=DATE-TIME:20220916T080000Z DTEND;VALUE=DATE-TIME:20220916T090000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10049@ictp.it DESCRIPTION:\n Membrane proteins play abundant roles in cellular and trans port functions\, thus represent a significant class of drug targets. Howev er\, their extraction and in vitro investigations are challenging due to t heir very hydrophobic nature.  Conventionally\, the extraction is achieve d with the aid of a surface-active compound known as detergent\, but most detergents cause irreversible loss of structure and activity to membrane p roteins. To solve these problems scaffold proteins or amphiphilic polymers such as styrene/maleic acid (SMA) copolymers were introduced as alternati ves to detergent\, but the lipid order was affected and there were precip itations in the presence of divalent cations. We show that an alternating diisobutylene/maleic acid (DIBMA) copolymer shows equal performance to SMA in solubilizing phospholipids\, stabilizes an integral membrane enzyme in functional bilayer nanodiscs\, and extracts proteins of various sizes dir ectly from cellular membranes. Unlike aromatic SMA\, aliphatic DIBMA has o nly a mild effect on lipid acyl-chain order\, does not interfere with opti cal spectroscopy in the far-UV range\, and does not precipitate in the pre sence of low millimolar concentrations of divalent cations. DIBMA was foun d to be highly charged and had effect on the electrostatic environment of the membrane proteins\, hence the development of other polymers such of su pho-DIBMA and Glyco-DIBMA. \n\n//indico.ictp.it/event/10049/ LOCATION:ICTP Common area\, second floor\, former SISSA building URL://indico.ictp.it/event/10049/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Games\, Graphs\, and Germs DTSTART;VALUE=DATE-TIME:20221026T090000Z DTEND;VALUE=DATE-TIME:20221026T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10098@ictp.it DESCRIPTION:\n This talk discusses combining experimental and theoretical approaches (drawn from graph theory and game theory) to study microbes in diverse scenarios.\n First\, we present a general\, rigorous mathematical framework to address mutagenesis -- from the level of a single protein to the organism level. We theoretically test the effectiveness of our measure in diverse microbial species to capture the reported effects of well-stud ied mutations and predict new ones. We further examine the usefulness of o ur procedure on a mutant of Mycobacterium smegmatis mc2155\, generated in our lab\, which is resistant to the mycobacteriophage D29 and exhibits sig nificant phenotypic alterations [1]. \n Using simple game theoretical mod eling\, we then show how mass testing and proactiveness affect epidemic sp reading [2].\n Along the way\, we also show that games on scale-free netwo rks may not witness cooperation to the wide extent\, as commonly thought [ 3].\n References:\n [1] Bioinformatics [Oxford]\,  37\,  213-220 (2021)  \n [2] J. Indian Institute of Science [Springer]\, 101\, 371-380 (2021)\ n [3] Europhysics Letters [EPL]\, 134\, 60002 (2021)\n\n//indico.ictp.it/e vent/10098/ LOCATION:ICTP Common area\, secondo floor\, old SISSA building URL://indico.ictp.it/event/10098/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Multiscale Approach in Physical Ecology of Insects: Modelling of Aerosols-carrying Pheromones DTSTART;VALUE=DATE-TIME:20221123T130000Z DTEND;VALUE=DATE-TIME:20221123T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10137@ictp.it DESCRIPTION:\n Chemical communication is ubiquitous in insects and plants life\, but complete understanding of the paths taken by chemicals to expla in the performance of these systems is not achieved. Critical gaps exist i n the understanding of the physico-chemistry of these chemicals for their transport in the near-ground atmosphere. We proposed that some of them for m films on aerosol surface and get transported as clusters. Therefore\, ae rosols could carry chemical signals over long distances.\n     In this talk\, we will show that mesoscale modelling tells that the kinetic of thi s adsorption process is sufficiently fast to occur during the atmospheric transport of the pheromones.  To assess this mechanism\, identification o f molecular candidates and model interface leading to high enough adsorpti on Gibbs energy is necessary. We will further present molecular dynamics s imulations that enable to compare\, at the microscale\, the film formation and 2D phase state of an iconic pheromone (bombykol) and three of its der ivatives on pure water.  These results\, and their multiscale integration \, help us to gain insights in the processes that play a role in pheromona l communication of insects and illustrates the importance of physico-chemi stry in environmental science.\n\n \n  \n\n//indico.ictp.it/event/10137/ LOCATION:ICTP Common area\, second floor\, old SISSA building URL://indico.ictp.it/event/10137/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Stochasticity in cellular processes DTSTART;VALUE=DATE-TIME:20221124T100000Z DTEND;VALUE=DATE-TIME:20221124T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10140@ictp.it DESCRIPTION:\n All cellular processes are characterized by stochasticity ( or noise) and the kind of noise involved has profound effects on their dyn amics. In this talk\, I will discuss the effects of noise on two specific biological processes viz. (i) mRNA translation process and (ii) virus-immu ne response pathway.\n  \n In the first problem we model extrinsic noise in mRNA translation\, and find\, among other effects\, the residence time distribution broadens significantly when the noise influences the entry ra te (initiation rate of ribosome on the mRNA) in comparison to the exit rat e of ribosomes. The ribosome residence time distribution obtained due to t he presence of extrinsic noise in the translation initiation rate is also similar to that obtained for the natural codon specific variation of initi ation rates [1].\n  \n In the second problem\, I will discuss our model o f the interaction between T cells (immune cells) and viruses. We introduce d fractional Gaussian noise in order to account for variability in the dyn amics of the entities and obtain an analytical expression for the joint pr obability density function of these variables and the average viral load i n the early and late stages of infection [2]. In further extensions\, we d eveloped and studied a more complex model through simulations [3]. Upon co mparing the theoretically predicted average virus levels to those of COVID -19 patients\, we hypothesize that the long-lived dynamics that are charac teristic of such viral infections are due to the long range correlations i n the temporal fluctuations of the virions.\n  \n The effects of noise on these specific cellular processes illustrate the importance of studying v ariability in these systems.\n  \n  \n References:\n  \n [1] Extrinsic noise effects on ribosomal traffic during the translation process. Rati Sh arma*. J. Stat. Mech: Theory Exp. 053504 (2022).\n [2] A near analytic sol ution of a stochastic immune response model considering variability in vir us and T cell dynamics. Abhilasha Batra and Rati Sharma*. J. Chem. Phys. 1 54\, 195104 (2021).\n [3] Persistent correlation in cellular noise determi nes longevity of viral infections. Abhilasha Batra\, Shoubhik Chandan Bane rjee and Rati Sharma*. J. Phys. Chem. Lett. 13\, 7252 (2022).\n\n//indico. ictp.it/event/10140/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10140/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Physics-Informed Data-Driven Tools: From Ideal models to Geophysical systems DTSTART;VALUE=DATE-TIME:20221129T100000Z DTEND;VALUE=DATE-TIME:20221129T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10130@ictp.it DESCRIPTION:\n \n \n Our ability to collect data is rapidly increasing t hanks to computational power and the unprecedented diversity of sensors. B ut how good are we at extracting\, reconstructing\, and understanding info rmation from them? Big data are not easy to digest. Our knowledge about th e Earth system is not increasing at the same speed as data\, but it follow s the slower progress of Data Reconstruction and Modelling techniques. Per forming in-silico experiments on ideal/complex systems\, we examine the ap plicability of Physics-Informed Data-Driven (PIDD) tools to extract interp retable information from sparse & heterogeneous observations of complex fl ows. Key applications are data-driven physics modeling\, data augmentation \, and super-resolution. We compare state-of-the-art purely Data-Driven to ols\, such as Generative Adversarial Networks (GAN) [1]\, with purely Phys ics-Informed approaches\, such as Nudging [2]\, on the ability to reconstr uct missing data. At the same time\, we will discuss how the potentials of both techniques are combined in the most advanced Physics-Informed Neural Networks. Moving from ideal cases such as turbulence on a rotating frame or Rayleigh-Benard convection I will discuss how PIDD tools can potentiall y result in a ground-breaking methodological revolution of today's data an alysis and data-reconstruction techniques in applications such as ocean ob servations [3\,4]. \n  \n\n\n [1] M. Buzzicotti\, F. Bonaccorso\, P. Cl ark Di Leoni\, L. Biferale. Reconstruction of turbulent data with deep gen erative models for semantic inpainting from TURB-Rot database. Physical Re view Fluids 6 (5)\, 050503\, (2021).\n\n [2] M. Buzzicotti\, P. Clark Di L eoni. Synchronizing subgrid scale models of turbulence to data. Physics of Fluids 32 (12)\, 125116\, (2021).\n\n [3] B.A. Storer\, M. Buzzicotti\, H . Khatri\, S.M. Griffies\, H. Aluie. Global energy spectrum of the general oceanic circulation. Nature communications 13 (1)\, 1-9\, (2022).\n [4] M . Buzzicotti\, B.A. Storer\, S.M. Griffies\, H. Aluie. A coarse-grained de composition of surface geostrophic kinetic energy in the global ocean. Ear th and Space Science Open Archive: ESSOAr  & arXiv preprint arXiv:2106.04 157\, (2021).\n\n//indico.ictp.it/event/10130/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10130/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: A macroecological law of interaction in microbial eco systems DTSTART;VALUE=DATE-TIME:20230112T100000Z DTEND;VALUE=DATE-TIME:20230112T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10235@ictp.it DESCRIPTION:\n Microbial communities are found all across the biosphere\, from human guts to glaciers\, from soil to activated sludge. Understanding common properties of such diverse communities can pave the way to elucida te the common mechanisms behind their patterns of correlations\, stability \, and resilience. In particular\, shedding light on how bacteria interact as a function of their genetic similarity is extremely relevant both at t he fundamental and practical levels. Using data from natural communities a nd mathematical modelling\, we discover  a macroecological law relating m ean correlation with genetic similarity\, revealing that correlation goes from positive to null values by increasing species dissimilarity. Fluctuat ions of shared environmental factors\, such as temperature or resources\, are responsible for such a universal pattern.\n\n//indico.ictp.it/event/10 235/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10235/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Hysteresis cycle and entropy production in a double-w ell potential DTSTART;VALUE=DATE-TIME:20230419T120000Z DTEND;VALUE=DATE-TIME:20230419T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10334@ictp.it DESCRIPTION:\n It was recently shown experimentally that resetting a singl e bit of information can be achieved at a work cost lower than Landauer's limit: k_B T ln(2)\, where T is the temperature of the environment [1]. Th is was achieved by generating a non-equilibrium steady state for the memor y\, itself encoded in the position of a particle jumping stochastically be tween the two wells of a double-well potential. When this double well pote ntial is made of two adjacent harmonic potentials\, the jump rates can be analytically obtained and the entropy production of each trajectory estima ted. This allows for the quantification of the total entropy production ge nerated during the erasure protocol.\n In this seminar\, I explain how thi s can be done and show experimental plots and preliminary results obtained in the case of an hysteresis cycle in the memory state.\n I will then pre sent new analytical results regarding gates performed on a massive flying particle.\n [1] Stochastic thermodynamics : driving of micro-oscillators a pplied to the study and the optimization of information processing by Sala mbô Dago\, Chapter 5\, https://theses.hal.science/tel-03771837v1/document \n\n//indico.ictp.it/event/10334/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/10334/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Analyzing Football as a Complex System DTSTART;VALUE=DATE-TIME:20230509T090000Z DTEND;VALUE=DATE-TIME:20230509T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10354@ictp.it DESCRIPTION:\n In this talk\, we present a study on the dynamics of markin g in football matches. To do this\, we survey and analyze a database conta ining the trajectories of players from both teams on the field of play dur ing three professional games. We describe the dynamics through the constru ction of temporal bipartite networks of proximity. Based on the introduced concept of proximity\, the nodes are the players\, and the links are defi ned between opponents that are close enough to each other at a given momen t. By studying the evolution of the heterogeneity parameter of the network s during the game\, we characterize a scaling law for the average shape of the fluctuations\, unveiling the emergence of complexity in the system. M oreover\, we propose a simple model to simulate the players’ motion in t he field from where we obtained the evolution of a synthetic proximity net work. We show that the model captures with a remarkable agreement the comp lexity of the empirical case\, hence it proves to be helpful to elucidate the underlying mechanisms responsible for the observed phenomena.\n\n//ind ico.ictp.it/event/10354/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/10354/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Discovery of 14-3-3 protein allosterisms by molecular dynamics DTSTART;VALUE=DATE-TIME:20230512T070000Z DTEND;VALUE=DATE-TIME:20230512T080000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10357@ictp.it DESCRIPTION:\n Molecular dynamics techniques can help unravel the function s of biomolecules in many ways. In this talk\, I will describe how we disc overed a completely new allosteric binding site in a protein family called 14-3-3. This protein family is involved in a fundamental mechanism of pro tein phosphorylation\, a mechanism used in eukaryotic cells to activate or inactivate thousands of proteins.\n Using molecular dynamics\, we were ab le to discover the allosteric site\, find and improve an active agent for this site\, and describe the subsequent step after binding to the final co nformational change in 14-3-3 that closes (and inactivates) the protein.\n \n//indico.ictp.it/event/10357/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/10357/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Self-propulsion induced aggregation in a binary mixtu re of active Brownian patchy-particles DTSTART;VALUE=DATE-TIME:20230526T090000Z DTEND;VALUE=DATE-TIME:20230526T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10379@ictp.it DESCRIPTION:\n Active matter systems\, which may involve the study of the combined actions of self-propelled particles\, have gained attention as a novel research field with widespread applications. Previous investigations have highlighted that the presence of self-propulsion is responsible for the emergence of collective phenomena among interacting particles\, phenom ena that do not arise when particles solely rely on thermal and diffusive effects for their motion. In this research\, we demonstrate the influence of self-propulsion on the formation of clusters in a binary mixture of act ive patchy-particles. These particles interact through both attractive and repulsive short-ranged forces. Gaining a deeper comprehension of such col lective behaviors\, along with others observed in active matter\, holds th e potential for future technological advancements.\n\n//indico.ictp.it/eve nt/10379/ LOCATION:ICTP Galileo Guest House - Fibonancci Room URL://indico.ictp.it/event/10379/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Biotremology: who we are\, where we go? DTSTART;VALUE=DATE-TIME:20230605T090000Z DTEND;VALUE=DATE-TIME:20230605T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10377@ictp.it DESCRIPTION:\n The increased awareness that substrate vibrations are an an cient and widespread form of animal communication\, as well as that vibrat ion receptors are ubiquitous in organisms\, lead to the establishment of b iotremology as discipline of study of vibrational behaviour. Following off icial recognition\, the scope of biotremological studies increased enormou sly.\n The evolution of vibrational communication did not involve only an emitter-receiver dyad and did not take place on the vibration-damped table in a sound-proof room in the laboratory. It took place in a complex natur al vibrational environment that is hidden to unaided human senses and view ing vibrational communication in a more ecological context open many impor tant research questions that will change our perception and understanding of the world around us.\n\n//indico.ictp.it/event/10377/ LOCATION:ICTP Galileo Guest House - Fibonacci room URL://indico.ictp.it/event/10377/ END:VEVENT BEGIN:VEVENT SUMMARY:Joint ICTP-SISSA Seminar: Stochastic thermodynamics of Boolean cir cuits\, finite automata and Turing machines DTSTART;VALUE=DATE-TIME:20230606T120000Z DTEND;VALUE=DATE-TIME:20230606T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10386@ictp.it DESCRIPTION:\n The central concern of computational complexity theory is t he minimal "resource costs" needed to perform a given computation on a giv en type of computer. In the real world\, some of the most important resour ce costs of performing a computation are thermodynamic\, e.g.\, the amount of heat it produces. In this talk I will summarize recent results on how thermodynamic resource costs depend on the computation being performed and the computer being used to perform them. I will start with some new resul ts concerning the thermodynamic costs of performing a given computation in a (loop-free and branch-free) digital circuit. \n\n Next I will summariz e some results concerning deterministic finite automata (DFA). After that I will review results on how considering the minimal entropy production (E P) of computing a desired output on a TM\, rather than the minimal size of an input string that causes the TM to produce that output (i.e.\, the out put's Kolmogorov complexity)\, results in a correction term to Kolmogorov complexity. I will end by describing the vast new set of research issues a t the intersection ofstochastic thermodynamics and computer science theory \, issues that expand both fields.\n\n//indico.ictp.it/event/10386/ LOCATION:SISSA SISSA via Bonomea 265\, Room 131 URL://indico.ictp.it/event/10386/ END:VEVENT BEGIN:VEVENT SUMMARY:Joint CMSP-QLS seminar: Multiple phase transitions in spin-1 model s with quenched random fields DTSTART;VALUE=DATE-TIME:20230612T090000Z DTEND;VALUE=DATE-TIME:20230612T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10392@ictp.it DESCRIPTION:\n In this talk I will discuss two versions of the famous Blum e Capel model\, one with random crystal field disorder and another with ra ndom magnetic field disorder for infinite range interactions.\n I will int roduce a method based on the theory of large deviations to perform the dis order averaging. We find novel phase diagrams with new phases and multi-cr itical points like the critical end point\, bicritical point and tetra cri tical point. I will also discuss a generalisation of the Blume-Capel model to p-spin interaction. This model with random crystal field disorder has a Gardner-like transition for finite p>=3.\n\n//indico.ictp.it/event/10392 / LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/10392/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Dipolar density inter-domain interaction in lipid mon olayers DTSTART;VALUE=DATE-TIME:20230621T090000Z DTEND;VALUE=DATE-TIME:20230621T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10406@ictp.it DESCRIPTION:\n A challenging problem in the biophysics of cell membranes i s the characterization of the lateral organization of their domains. These domains are made up of lipids in an ordered phase and are dispersed in a more disordered continuous lipid phase. A widely used experimental model t o study these systems are the lipid monolayers at the water-air interface. Based on these experiments and using Brownian simulations\, we study how the dipolar interaction between domains affects the structure of monolayer s with phase coexistence. We present a new way to estimate the dipole repu lsion directly from the experimental results. Besides\, we also study an a nalytical approach to this system based on the Ornstein-Zernike integral e quation theories (OZ-IETs). When the pair correlation function serves as t he measure\, we found that the closures that outperform better are the Mod ified Hypernetted Chain (MHNC) and the Variational Modified Hypernetted Ch ain (VMHNC)\, with a reference system based on an extension of the hard sp heres parametrization. On the other hand\, when the structure factor is us ed as a measure\, the reference system that performs better depends on th e particular region of the phase diagram.\n\n//indico.ictp.it/event/10406/ LOCATION:ICTP Galileo Guest House - Fibonacci Room URL://indico.ictp.it/event/10406/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Dynamics of active nematic defects on cones DTSTART;VALUE=DATE-TIME:20230623T090000Z DTEND;VALUE=DATE-TIME:20230623T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10389@ictp.it DESCRIPTION:\n In the first part of the talk\, we investigate the ground-s tate configurations of two-dimensional liquid crystals with p-fold rotatio nal symmetry (p-atics) on cones. The cone apex develops an effective topol ogical charge\, which in analogy to electrostatics\, leads to defect absor ption and emission at the cone apex as the deficit angle of the cone is va ried. We find three types of ground-state configurations as a function of cone angle\, which is determined by charged defects screening the effectiv e apex charge: (i) for sharp cones\, all of the +1/p defects are absorbed by the apex\; (ii) at intermediate cone angles\, some of the +1/p defects are absorbed by the apex and the rest lie equally spaced along a concentri c ring on the flank\; and (iii) for nearly flat cones\, all of the +1/p de fects lie equally spaced along a concentric ring on the flank. We check th ese results with numerical simulations for a set of commensurate cone angl es and find excellent agreement. In the second part of the talk\, we inves tigate the dynamics of an active nematic on a cone\, and via simulations f ind long-time circular orbits of either one or two flank defects\, with tr ansitions between these states mediated by the apex via defect absorption\ , emission\, or defect pair creation. \n\n//indico.ictp.it/event/10389/ LOCATION:ICTP Galileo Guest House - Fibonacci Room URL://indico.ictp.it/event/10389/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Multiscale Simulations for an affordable description of virus in motion: Symmetry-related dynamics determines ZIKA and Dengue n eutralization DTSTART;VALUE=DATE-TIME:20230626T090000Z DTEND;VALUE=DATE-TIME:20230626T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10410@ictp.it DESCRIPTION:\n \n \n Molecular dynamics simulation running simultaneous molecular granularities drastically reduce computational costs\, making t he study of whole viruses available to desktop computers. I will illustrat e the use of a coarse-grained force field for biomolecules to perform a se t of simulations on Zika\, Dengue\, and other related viruses. Comparativ e analysis among five common viruses pinpoints a collective variable that shows similar behavior only in the 3-fold symmetry axis of the icosahedral structure and coincides with the binding site of broadly neutralizing ant ibodies for Zika and Dengue.\n Noteworthy\, this dynamical similarity is only conserved in the same symmetry region of phylogenetically close viru ses\, suggesting an Achilles heel for crossed  neutralization of similar viruses.\n \n \n \n  \n \n \n\n\n  \n\n//indico.ictp.it/event /10410/ LOCATION:ICTP Galileo Guest House - Fibonacci room URL://indico.ictp.it/event/10410/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Learning without detailed balance DTSTART;VALUE=DATE-TIME:20230627T090000Z DTEND;VALUE=DATE-TIME:20230627T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10380@ictp.it DESCRIPTION:\n The assumption of rational traders has been the subject of a harsh debate in theoretical economics for the past 40 years. Recent expe rimental evidence gathered by cognitive neuroscientists suggests that the way we learn in simple tasks is in stark contrast with the rationality ass umption: the way we learn is biased. Confirmation bias\, for example\,\n i s the tendency to incorporate the information in line with our priors and disregard the information in contrast with them.\n These findings beg for an explanation. In particular\, if evolution selected such biases\, they s hould be beneficial in some circumstances: for example\, in tasks with two asymmetric bandits\, there are 'optimal biases' which allow individuals t o increase the average earned reward. The reason for these additional gain s is that biased beliefs are magnified\, allowing one to distinguish more clearly between two similar options.\n Interestingly\, in these contexts\, the optimal bias corresponds to a learning dynamics that breaks detailed balance\, leading\, therefore\, to irreversible dynamics\, even in the sta tionary state. We argue\, by means of analytical calculations and numerica l simulations\, that the optimal bias corresponds to dynamics that maximiz e the entropy production in the stationary state. In particular\, in this context\, the stationary state with maximum entropy production allows agen ts to safely explore the environment without being stuck in a sub-optimal belief.\n\n//indico.ictp.it/event/10380/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/10380/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Computational genomics and dark DNA: data types\, app roaches\, methods DTSTART;VALUE=DATE-TIME:20230630T090000Z DTEND;VALUE=DATE-TIME:20230630T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10417@ictp.it DESCRIPTION:\n Abstract:\n Computational genomics is a field of research t hat utilizes robust computational and statistical approaches to decode the functional information hidden in genomic and transcriptomic sequences. Dr astically lowered sequencing costs have resulted in accumulation of massiv e amounts of genomic datasets with estimates that predict that genomics re search will generate between 2 and 40 exabytes of data within this decade. Our ability to fully decipher knowledge that sequencing datasets contain is far behind the advancements of sequencing techniques. This makes comput ational genomics a vibrant field of research for the foreseeable future.\n  \n As is well-known\, DNA provides the necessary instructions for the c reation and maintenance of the proteins responsible for constructing and s upporting our bodies. Historically\, most of the effort was focused on stu dying this protein-coding part of the genome. However\, only about 2% of o ur DNA encodes proteins. The remaining 98% is non-coding DNA which was lon g regarded as “junk” DNA. Over time\, scientific studies have revealed the significant role of non-coding DNA in both biological processes and t he process of evolution.\n  \n In the first part of the talk\, I will giv e an introduction to the field of computational genomics highlighting the most common approaches and data types. In the second part\, I will focus o n the “dark” DNA\, namely on transposable elements: mobile genetic ele ments that are able to self-propagate\, and represent almost half of the h uman genome. I will show by several examples how computational approaches can facilitate scientific discovery in biology.\n  \n\n//indico.ictp.it/e vent/10417/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/10417/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: System 2 Applied Mathematics DTSTART;VALUE=DATE-TIME:20230704T090000Z DTEND;VALUE=DATE-TIME:20230704T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10412@ictp.it DESCRIPTION:According to psychologist D. Kahneman (2002 Nobel prize winner in economics for the work on decision making under uncertainty) there are two modes/systems of thinking. System 1 operates automatically and quickl y\, like deep learning empowered by automatic differentiation.  System 2 allocates attention to the effortful mental activities\, like building exp lainable heuristics in quantitative sciences.\n \nIn this talk we illus trate how applied mathematics is harnessing system 1 achievements of moder n AI to build system 2 for quantitative sciences. Specifically\, we discus s design\, training and validation of the system 2 for\n \n\n 1.  Physic s Informed Machine Learning of Power Systems\n 2.  Lagrangian Large Eddy Simulations of Turbulent Flows\n 3.  Graphical and Agent Based Models of epidemiological bursts\n \n \n Bio: Dr. Michael (Misha) Chertkov is a Professor of Mathematics and Chair of the Graduate Interdisciplinary Prog ram (GIDP) in Applied Mathematics at the University of Arizona (UArizona). Dr. Chertkov area of focus is mathematics\, including statistics and da ta science\, applied to physical\, engineered and other systems and networ ks. He  has published more than 250 papers\, is a fellow of AAAS\, a fell ow of the American Physical Society and a senior member of IEEE.\n\n//indi co.ictp.it/event/10412/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/10412/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Novel Dynamical Models of Wealth and Cooperation Usin g Information Theory DTSTART;VALUE=DATE-TIME:20230712T080000Z DTEND;VALUE=DATE-TIME:20230712T090000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10422@ictp.it DESCRIPTION:\n \n The statistical dynamics of population growth\, wealth \, and inequality are related through their common description as multipli cative stochastic processes. In these systems\, variations in growth rates determine selection\, relative wealth\, and cooperation. Despite its re levance for biological and social sciences\, we still lack a general theo ry that explains the dynamics of growth rates in terms of agent adaptatio n to their environment\, heterogeneities of traits\, and other behaviors. In this talk\, we derive a general population dynamics of learning agents sharing and leveraging the knowledge of their environments to grow and rei nvest their resources as individuals or groups. Growth rates emerge in the long-time limit as the mutual information between agents' signals and s tates of the environment and information synergies across signals. We show that\, with knowledge of past and present conditions\, sequential Bayesia n inference becomes optimal for maximizing growth\, formally associating g rowth rate dynamics with information dynamics. We discuss how this framew ork can address problems of inequality and cooperation in heterogeneous po pulations and lay the foundations for more robust interacting growth mode ls\n\n\n  \n\n//indico.ictp.it/event/10422/ LOCATION:ICTP Common area\, old SISSA building URL://indico.ictp.it/event/10422/ END:VEVENT BEGIN:VEVENT SUMMARY:Joint CMSP-QLS seminar: Quantum unitary evolution interspersed wit h repeated non-unitary interactions at random times DTSTART;VALUE=DATE-TIME:20230724T120000Z DTEND;VALUE=DATE-TIME:20230724T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10424@ictp.it DESCRIPTION:\n We address the issue of what happens when the unitary evolu tion of a generic closed quantum system is interrupted at random times wit h non-unitary evolution due to interactions with either the external envir onment or a measuring apparatus. We adduce a general theoretical framework to obtain the average density operator of the system at any time during t he dynamical evolution\, which is applicable to any form of non-unitary in teraction. We provide two explicit applications of the formalism in the co ntext of the so-called tight-binding model relevant in various contexts in solid-state physics\, for two representative forms of interactions: (i) s tochastic resets\, whereby the density operator is at random times reset t o its initial form\, and (ii) projective measurements at random times. For (i)\, we demonstrate with our exact results how the particle is localized on the sites at long times\, leading to a time-independent mean-squared d isplacement of the particle about its initial location. For (ii)\, we show that repeated projection to the initial state of the particle results in an effective suppression of the temporal decay in the probability of the p article to be found on the initial state. The amount of suppression is com parable to the one in conventional Zeno effect scenarios\, but which does not require to perform measurements at exactly regular intervals that are hallmarks of such scenarios.\n  \n\n//indico.ictp.it/event/10424/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/10424/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Effects of linking topology on the shear response of connected ring polymers: Catenanes and bonded rings flow differently DTSTART;VALUE=DATE-TIME:20230724T090000Z DTEND;VALUE=DATE-TIME:20230724T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10426@ictp.it DESCRIPTION:\n Over the last two decades\, there has been a growing intere st in studying polymers with different architectures in shear flow. Theref ore\, there is a vast body of literature on the relationship between topol ogy\, hydrodynamic interaction\, and shear flow.\n In recent times\, signi ficant progress has been made in the synthesis of poly[n]catenanes. These systems possess mechanical bonds that introduce fresh degrees of freedom\, opening up avenues for material engineering and new possibilities in dyna mics.\n In this talk\, I aim to explain the behavior of poly[2]catenane wh en it is out of equilibrium. Our objective is to gain insight into the nat ure of mechanical bonds in polymers and the impact of hydrodynamic interac tion in systems comprising two rings.\n Authors:\n Reyhaneh Afghahi Farima ni (Sharif University of Technology\, University of Vienna)\n Zahra Ahmadi an Dehaghani (SISSA)\n Christos Likos ( University of Vienna)\n Reza Ejteh adi (Sharif University of Technology)\n\n//indico.ictp.it/event/10426/ LOCATION:ICTP Common area\, ex SISSA building URL://indico.ictp.it/event/10426/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Entropy and statistical complexity rates of Student's t-process DTSTART;VALUE=DATE-TIME:20230906T120000Z DTEND;VALUE=DATE-TIME:20230906T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10528@ictp.it DESCRIPTION:Abstract:\n Student's t-process is a parametrized generalizati on of the Gaussian process\, which is widely used for the analysis of heav y-tailed data. Although these processes share several principal properties \, an important difference can be observed once their Renyi entropy rates are analyzed instead of the commonly considered Shannon entropy rate. In t his talk\, we show that stationary Student's t-processes have\, in general \, a nonfinite Renyi entropy rate\, while a specific class of nonstationar y Student's t-processes has a finite Renyi entropy rate\, which is fully o pposed to the Shannon case. After that\, we extend considerations to the g eneralized statistical complexity rates\, which represent an interplay bet ween order and disorder levels of a random process. We show that the stati stical complexity rates are invariant with respect to the stationarity of the Student's t-process while keeping information about its shape paramete r.\n\n \n  \n\n//indico.ictp.it/event/10528/ LOCATION:ICTP Common Area\, ex SISSA building\, second floor URL://indico.ictp.it/event/10528/ END:VEVENT BEGIN:VEVENT SUMMARY:Information and thermodynamics: optimizing information processing using underdamped systems DTSTART;VALUE=DATE-TIME:20231003T120000Z DTEND;VALUE=DATE-TIME:20231003T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10547@ictp.it DESCRIPTION:\n The Landauer principle states that at least kBTln2 of energ y is required to erase a 1-bit memory\, with kBT the thermal energy of the system. Practical erasure implementations re- quire an overhead to Landau er’s Bound (LB)\, observed to scale as kBT × B/τ\, with τ the protoco l duration and B close to the system relaxation time. Most model experimen ts use overdamped systems\, for which minimizing the overhead means minimi zing the dissipation. Underdamped systems thus sound appealing to reduce t his energetic cost\, and are the object of this presentation. Our experime nt implements a model 1-bit memory based on a micro-mechanical oscillator confined in a double-well potential created by a feedback loop [1]. We mea sure the work and the heat of informa- tion processing protocols within th e stochastic thermodynamic framework. Our research covers all possible ope rations on a single bit b: HOLD (b → b)\, SET (b → 0 or 1)\, NOT (b → ¬b).\n  \n The logical SET operation is an erasure\, logically irrev ersible\, coming with an entropic cost which is at least LB. We demonstrat e that\, in our experiment\, this bound is reached with a 1% uncertainty\, with protocols as short as 100 ms [2]. Besides\, we show experimentally a nd theoretically that for underdamped systems\, fast erasures induce a hea ting of the memory: the work influx is not instantaneously compensated by the inefficient heat transfer to the thermostat. This temperature rise res ults in a kinetic and potential energy contribution superseding the viscou s dissipation term. Our model covering all damping regimes paves the way t o new optimization strategies in information processing [3\, 4]\, includin g the implementation of more applied logic gates performing repeated fast operations [6].\n  \n The other logical operations are reversible\, and c an thus in principle be performed at no ther- modynamical cost. We impleme nt the NOT operation using the momentum degree of freedom in our underdamp ed memory to perform a bit-flip [5]. Not bounded by any entropic cost this time\, the energetic cost of the protocol vanishes as the quality factor of the oscillator increases\, further highlighting the low energy footprin t and interest of underdamped memories.\n  \n References\n1. S. Dago\, J . Pereda\, S. Ciliberto and L. Bellon: JSTAT 5\, 053209 (2022)\n2. S. Dago \, J. Pereda\, N. Barros\, S. Ciliberto\, and L. Bellon: Phys. Rev. Lett. 126\, 17 (2021)\n3. S. Dago and L. Bellon: Phys. Rev. Lett. 128\, 7 (2022) \n4. S. Dago\, S. Ciliberto and L. Bellon: PNAS 120 (39) e2301742120\, (20 23).\n5. S. Dago and L. Bellon: Phys. Rev. E 108\, L022101\, (2023).\n6. S . Dago\, S.Ciliberto and L.Bellon: To be published in Advanced Physics Res earch\,\narXiv:2306.15573 (2023)\n\n//indico.ictp.it/event/10547/ LOCATION:ICTP Galileo Guest House - Fibonacci Room URL://indico.ictp.it/event/10547/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Mean-Field Analysis of Lasso under Ultra-Sparse Condi tions DTSTART;VALUE=DATE-TIME:20231025T120000Z DTEND;VALUE=DATE-TIME:20231025T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10566@ictp.it DESCRIPTION:\n We analyze the average performance of the least absolute sh rinkage and selection operator (Lasso) for the linear model under a Gaussi an matrix design\, when the number of regressors grows larger while keepin g the true support size finite\, i.e.\, the ultra-sparse case. The result is based on a novel treatment of the non-rigorous replica method in statis tical physics\, where self-averaging assumptions on certain random variabl es are relaxed. Using our theory\, the average performance for Gaussian me asurements and noise can be assessed from the solution of a random scalar optimization problem with O(1) random elements.\n\n//indico.ictp.it/event/ 10566/ LOCATION:ICTP Common Area\, Ex SISSA building\, second floor URL://indico.ictp.it/event/10566/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Non-equilibrium Dynamics of Fluids upon Temperature Q uench DTSTART;VALUE=DATE-TIME:20231108T130000Z DTEND;VALUE=DATE-TIME:20231108T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10578@ictp.it DESCRIPTION:\n When a fluid\, close to a second order phase transition\, i s quenched suddenly from a high temperature to a temperature below the cri tical point it falls out of equilibrium and proceeds towards the new equil ibrium state via a non-equilibrium dynamics known as phase separation. Pre sence of an inhomogeneous temperature field makes this dynamics more comp lex and interesting. In this direction\, results for a single-component fl uid will be presented from molecular dynamics simulations.\n\n//indico.ict p.it/event/10578/ LOCATION:ICTP Common Area\, Ex SISSA building\, second floor URL://indico.ictp.it/event/10578/ END:VEVENT BEGIN:VEVENT SUMMARY:Joint OGS-QLS Seminar: Diversity begets stability: sublinear growt h and competitive coexistence across ecosystems DTSTART;VALUE=DATE-TIME:20231122T093000Z DTEND;VALUE=DATE-TIME:20231122T103000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10588@ictp.it DESCRIPTION:\n The dramatic loss of species diversity brings urgency to un derstanding how diverse ecosystems are naturally stabilized. Whereas conve ntional wisdom and empirical observation suggest that\n stability increase s with diversity\, ecological theory has long made the opposite prediction \, leading\n to the longstanding “diversity-stability debate”. \n\n W e show this puzzle is resolved through a model where growth scales as a su blinear power law with biomass (exponent < 1)\, exhibiting a form of popul ation self-regulation analogous to models of individual ontogeny. We show that competitive interactions in a community with sublinear growth do not lead to exclusion\, but instead promote stability at higher diversity. \n \n Our model realigns theory with classic observations and predicts large- scale macroecological patterns. However\, it makes an unsettling predictio n: biodiversity loss may accelerate the destabilization of ecosystems.\n\n //indico.ictp.it/event/10588/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/10588/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: The ecology of the gut microbiome DTSTART;VALUE=DATE-TIME:20240122T100000Z DTEND;VALUE=DATE-TIME:20240122T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10621@ictp.it DESCRIPTION:\n Microbiota dynamics arise from a plethora of interspecies i nteractions\, including resource competition\, cross-feeding\, and pH modu lation. The individual contributions of these mechanisms are challenging t o untangle\, especially in natural or complex laboratory environments wher e the landscape of resource competition is unclear.\n First\, I will discu ss how we developed a framework to estimate the extent of multi-species ni che overlaps by combining metabolomics data of individual species\, growth measurements in pairwise spent media\, and mathematical models. When appl ied to an in vitro model system of human gut commensals in complex media\, our framework revealed that a simple model of resource competition descri bed most pairwise interactions. By grouping metabolomic features depleted by the same set of species\, we constructed a coarse-grained consumer-reso urce model that predicted assembly compositions to reasonable accuracy.\n Next\, I will show how dose dependence can arise between closely related o rganisms as long as competition is sufficiently intense. Finally\, I will discuss our recent efforts to explore spatial variation along the human in testinal tract\, and how that informs our understanding of the gut environ ment.\n  \n\n//indico.ictp.it/event/10621/ LOCATION:ICTP Common Area\, Ex SISSA Building URL://indico.ictp.it/event/10621/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Fast renormalizing the structures and dynamics of ult ra-large systems via random renormalization group DTSTART;VALUE=DATE-TIME:20240228T090000Z DTEND;VALUE=DATE-TIME:20240228T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10661@ictp.it DESCRIPTION:\n Criticality and symmetry\, studied by the renormalization g roups\, lie at the heart of modern physics theories of matters and complex systems. However\, surveying these properties with massive experimental d ata is bottlenecked by the intolerable costs of computing renormalization groups on real systems. Here\, we develop a time- and memory-efficient fra mework\, termed as the random renormalization group\, for renormalizing ul tra-large systems (e.g.\, with millions of units) within minutes. This fra mework is based on random projections\, hashing techniques\, and kernel re presentations\, which support the renormalization governed by linear and n on-linear correlations. For system structures\, it exploits the correlatio ns among local topology in kernel spaces to unfold the connectivity of uni ts\, identify intrinsic system scales\, and verify the existences of symme tries under scale transformation. For system dynamics\, it renormalizes un its into correlated clusters to analyze scaling behaviours\, validate scal ing relations\, and investigate potential criticality. Benefiting from has hing-function-based designs\, our framework significantly reduces computat ional complexity compared with classic renormalization groups\, realizing a single-step acceleration of two orders of magnitude. Meanwhile\, the eff icient representation of different kinds of correlations in kernel spaces realized by random projections ensures the capacity of our framework to ca pture diverse unit relations. As shown by our experiments\, the random ren ormalization group helps identify non-equilibrium phase transitions\, crit icality\, and symmetry in diverse large-scale genetic\, neural\, material\ , social\, and cosmological systems.\n  \n\n//indico.ictp.it/event/10661/ LOCATION:ICTP Common Area\, ex SISSA building URL://indico.ictp.it/event/10661/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Information propagation across timescales DTSTART;VALUE=DATE-TIME:20240404T120000Z DTEND;VALUE=DATE-TIME:20240404T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10677@ictp.it DESCRIPTION:\n The presence of interconnected processes taking place at mu ltiple temporal scales is one of the fundamental features of biochemistry\ , neural networks\, ecological communities\, and many other complex system s. A key feature is that these processes may interact both directly and in directly with one another\, and such interactions across timescales may no t be pairwise. This makes understanding the relationship between their com ponents a formidable challenge. In this talk\, we will first focus on how the different timescales associated with each layer impact their effective couplings. We derive a general decomposition of the joint probability dis tribution that is valid for any set of dynamical processes\, and we introd uce the general principles governing how information propagates across tim escales. In doing so\, we elucidate the interplay between mutual informati on and causality in multiscale systems. We show how this information struc ture can exploited to study minimal models of biochemical signaling networ ks and to understand the relations between their fundamental components. F urther\, we apply our framework to characterize the effective dependencies induced by changing environmental conditions in coupled stochastic proces ses. Finally\, we outline a promising future perspective where we introduc e the concept of information precision\, which incorporates the variance o f mutual information along stochastic trajectories.\n\n//indico.ictp.it/ev ent/10677/ LOCATION:ICTP Common Area\, ex SISSA building\, second floor URL://indico.ictp.it/event/10677/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Where to locate the beehives and bee pasture? DTSTART;VALUE=DATE-TIME:20240410T120000Z DTEND;VALUE=DATE-TIME:20240410T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10688@ictp.it DESCRIPTION:\n In this presentation\, I will introduce a nuanced approach aimed at curbing the transmission of infectious diseases among bees\, a cr ucial consideration given their role as pollinators for numerous plant spe cies. While bees serve as essential contributors to ecosystems\, they are susceptible to microbial diseases\, which can spread within their communit ies. An infected bee traversing its foraging area may inadvertently leave behind infective spores\, posing a risk to other bees and the plants they visit. To address this challenge\, one proposed solution involves beekeepe rs implementing a buffer foraging area and strategically locating beehives to segregate bee colonies. However\, this strategy presents a dual potent ial: it can serve as a protective barrier\, mitigating disease transmissio n\, or it may inadvertently attract more bees\, thereby increasing the lik elihood of infection spread. In this discussion\, we will delve into the i ntricate dynamics underlying this strategy using mathematical modeling. We will explore factors such as the impact of foraging distances\, the influ ence of stochastic bee behavior\, the implications of optimal foraging the ory\, and strategies for designing the buffer area to minimize negative ou tcomes. Through this analysis\, we aim to offer insights into optimizing b eekeeping practices to safeguard bee health and enhance ecological resilie nce. \n  \nReferences[1] M.K.A. Gavina\, J.F. Rabajante and C.R. Cervanc ia\, Mathematical programming models for determining the optimal location of beehives\, Bulletin of Mathematical Biology\, 76 (997-1016)\, 2014.[2] E.O. Jatulan\, J.F. Rabajante\, C.G.B. Banaay\, A.C. Fajardo Jr. and E.C. Jose\, A mathematical model of intra-colony spread of American Foulbrood i n European honeybees (Apis mellifera L.)\, PLoS ONE\, 10 (e0143805)\, 2015 .[3] J.F. Rabajante\, J.M. Tubay\, E.C. Jose and C.R. Cervancia\, Pollinat or diversity and density measures: survey and indexing standard to model\, detect\, and assess pollinator deficits\, Modeling Earth Systems and Envi ronment\, 6 (363-371)\, 2020.\n\n//indico.ictp.it/event/10688/ LOCATION:ICTP Common area\, Ex SISSA building\, Second floor URL://indico.ictp.it/event/10688/ END:VEVENT BEGIN:VEVENT SUMMARY:Counterintuitive behavior in models of eco-evolutionary dynamics DTSTART;VALUE=DATE-TIME:20240521T120000Z DTEND;VALUE=DATE-TIME:20240521T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10713@ictp.it DESCRIPTION:\n In many large natural ecosystems\, the influx of mutations can be expected to be quite large (>>1 over the characteristic turnover ti me of the population).  Therefore\, we can say that\, in such ecosystems\ , the timescales of ecological and evolutionary change are not well separa ted.  Because of this lack of timescale separation\, it is possible for m utation\, selection and ecological dynamics to feedback on one another.  In this talk\, I will argue that common approaches to modeling the mutatio n process overlook known aspects of biology and thereby underestimate the impact of eco-evolutionary feedbacks.  I will propose a simple\, biologic ally motivated model of mutation that restores the effect of these feedbac ks.  When these mutational dynamics are coupled to an underlying model of ecology\, I will show that the long-term behavior of the ecosystem can be impacted in counterintuitive ways.  I will demonstrate this in the conte xt of two canonical classes of ecological models: 1) asymmetric predator-p rey models (an archetype of chaotic ecology) 2) consumer-resource models ( an archetype of stable ecology).  I will argue that similar eco-evolution ary feedbacks might be relevant in natural\, laboratory\, and human health -related ecosystems\, in particular in phage-bacteria communities and in t he clearance of pathogens by the adaptive immune system.\n\n//indico.ictp. it/event/10713/ LOCATION:ICTP Common Area\, ex SISSA building URL://indico.ictp.it/event/10713/ END:VEVENT BEGIN:VEVENT SUMMARY:Modeling Cryptochromes using Generative Models DTSTART;VALUE=DATE-TIME:20240611T120000Z DTEND;VALUE=DATE-TIME:20240611T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10734@ictp.it DESCRIPTION:\n Migration is an observed phenomenon in birds' life which is crucial to the species' survival [1\,2]. In several migrating avian speci es such as European blackcaps\, willow warbler\, or Eurasian cuckoo\, migr ation was found to have a genetic basis as an inherited trait that receive d selective evolutionary pressure rather than a learned behavior [3\,4]. S ome migratory birds use the inclination of the magnetic field of the earth as its magnetic compass for navigation [5].  The most promising candidat e that is responsible for the magnetoreception in avian navigation is the photoreceptor protein cryptochrome which is located in the outer segment o f the photoreceptor cells [6].\n  \n Inside the cryptochrome\, a radical pair [FAD∙-TrpH∙-] forms after a photon excites the FAD cofactor that is associated with the signaling of the protein. An electron transfer chai n is then formed among tryptophans to create a radical pair with FAD. The rates of the individual electron transfer processes affect the sensitivity of the species to the magnetic field [7].\n  \n Previous experiments hav e shown that different species have varying magnetic field sensitivity [8] . This makes it important to rationalize the rates of the electron transfe rs for each of the different cryptochromes in different migratory birds.\n  \n A common approach done is to model the protein in different redox st ates of the radical pair formation using molecular dynamics and estimate a ssociated the free energy [8\,9]. This method however is computationally t axing. In this talk\, we propose a method to use generative models to mode l cryptochrome using training data from molecular dynamics simulations.\n  \n We use the available data for cryptochrome from a European robin (ErC ry4) as a training set and generate different protein conformations using generative adversarial networks and variational autoencoders. From the tra ined model\, our goal is to generate protein conformations of cryptochrome of other species such as pigeon or chicken. The latent space distribution is modeled conditioned on the given protein structure of the European rob in so that training can be transferred to a new protein structure of anoth er species. This method can be used to quickly generate protein conformati ons such that energy and the electron transfer rates could be estimated fo r other cryptochrome structures such are those found in other migratory bi rds and in fishes.\n  \n  \n  \n [1]        Liedvogel\, M.\, Åke sson\, S.\, & Bensch\, S. (2011). The genetics of migration on the move. T rends in ecology & evolution\, 26(11)\, 561-569. https://doi.org/10.1016/j .tree.2011.07.009\n [2]        Dingle\, H.\, & Drake\, V. A. (2007). What is migration?. Bioscience\, 57(2)\, 113-121. https://doi.org/10.1641 /B570206\n [3]        Merlin\, C.\, & Liedvogel\, M. (2019). The gen etics and epigenetics of animal migration and orientation:birds\, butterfl ies and beyond. Journal of Experimental Biology\, 222(Suppl_1)\, jeb191890 . https://doi.org/10.1242/jeb.191890\n [4]        Lundberg\, M.\, Li edvogel\, M.\, Larson\, K.\, Sigeman\, H.\, Grahn\, M.\, Wright\, A.\, ... & Bensch\, S. (2017). Genetic differences between willow warbler migrator y phenotypes are few and cluster in large haplotype blocks. Evolution Lett ers\, 1(3)\, 155-168. https://doi.org/10.1002/evl3.15\n [5]        W iltschko\, W.\, Munro\, U.\, Ford\, H.\, & Wiltschko\, R. (1993). Magnetic inclination compass: a basis for the migratory orientation of birds in th e Northern and Southern Hemisphere. Experientia\, 49\, 167-170. https://do i.org/10.1007/BF01989423\n [6]        Ritz\, T.\, Adem\, S.\, & Schu lten\, K. (2000). A model for photoreceptor-based magnetoreception in bird s. Biophysical journal\, 78(2)\, 707-718. https://doi.org/10.1016/S0006-34 95(00)76629-X\n [7]        Timmel\, C. R.\, Till\, U.\, Brocklehurst \, B.\, Mclauchlan\, K. A.\, & Hore\, P. J. (1998). Effects of weak magnet ic fields on free radical recombination reactions. Molecular Physics\, 95( 1)\, 71-89. https://doi.org/10.1080/00268979809483134\n [8]        X u\, J.\, Jarocha\, L. E.\, Zollitsch\, T.\, Konowalczyk\, M.\, Henbest\, K . B.\, Richert\, S.\, ... & Hore\, P. J. (2021). Magnetic sensitivity of c ryptochrome 4 from a migratory songbird. Nature\, 594(7864)\, 535-540. htt ps://doi.org/10.1038/s41586-021-03618-9\n [9]        Schuhmann\, F.\ , Kattnig\, D. R.\, & Solov’yov\, I. A. (2021). Exploring post-activatio n conformational changes in pigeon cryptochrome 4. The Journal of Physical Chemistry B\, 125(34)\, 9652-9659. https://doi.org/10.1021/acs.jpcb.1c027 95\n  \n\n//indico.ictp.it/event/10734/ LOCATION:ICTP Common Area\, second floor\, ex SISSA buildin URL://indico.ictp.it/event/10734/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: An Uncertain Walk Through Games\, Conflicts and Netwo rks DTSTART;VALUE=DATE-TIME:20240625T090000Z DTEND;VALUE=DATE-TIME:20240625T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10751@ictp.it DESCRIPTION:\n In this talk\, I will describe part of the journey that too k me from researching modeling unawareness in games to studying applicati ons of graph models to analyze conflict and social/complex networks. The i dea of the talk is to give an overview of some of the work I have been doi ng throughout my career to pursue overlapping interests.\n\n//indico.ictp .it/event/10751/ LOCATION:ICTP Common area\, second floor\, ex SISSA building URL://indico.ictp.it/event/10751/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Hybrid Data Partitioning for Enhanced Machine Learnin g in High-Cost Systems DTSTART;VALUE=DATE-TIME:20240716T090000Z DTEND;VALUE=DATE-TIME:20240716T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10761@ictp.it DESCRIPTION:\n Effective data partitioning is critical in machine learning \, especially in high-cost physical systems. Cross- validation techniques like K-Fold Cross-Validation (KFCV)\, though used to enhance model robustn ess\, is known to compromise generalization assessment due to extensive co mputation and data shuffling demands. Simple Random Sampling (SRS) is used in this study to mitigate the weaknesses of KFCV. While SRS ensured repre sentative samples\, it risked non-representative sets in imbalanced data. Integrating both methods minimized biases\, blending the simplicity of the SRS with the accuracy of the KFCV to optimize data partitioning. The hybr id method enhanced mean and variance convergence across diverse dataset si zes and trials\, effectively balancing performance under various condition s. It proved beneficial in resource- constrained environments and with ext ensive datasets\, providing practical solutions for effective machine lear ning implementations.\n  \n\n//indico.ictp.it/event/10761/ LOCATION:ICTP Common area\, Ex SISSA building\, Second floor URL://indico.ictp.it/event/10761/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Chiral motion of a 3D model microswimmer DTSTART;VALUE=DATE-TIME:20240718T090000Z DTEND;VALUE=DATE-TIME:20240718T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10764@ictp.it DESCRIPTION:\n Generalizing a 2D micro-swimmer to 3D results in interestin g symmetry dependent swimming directions. It moves in helical paths\, whos e both chirality and the direction depend on  initial move. The swimmer a lso shows a perfect  ability to sense its environment and search for chem icals. \n\n//indico.ictp.it/event/10764/ LOCATION:ICTP Common Area\, ex SISSA building\, Second floor URL://indico.ictp.it/event/10764/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: Quantum computing with classical data: Beyond time co mplexity DTSTART;VALUE=DATE-TIME:20240729T080000Z DTEND;VALUE=DATE-TIME:20240729T090000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10767@ictp.it DESCRIPTION:\n If large\, fault-tolerant quantum computers are ever built\ , it is widely believed that they will greatly enhance our understanding o f quantum systems. \n\n However\, the impact of these computers on the pr ocessing of classical data remains uncertain. While there are a number of quantum algorithms offering potentially exponential speedups on genericall y useful classical tasks such as matrix inversion\, it is unclear if any s uch speedups hold when including the cost of moving classical data to and from a quantum computer\, and after carefully accounting for the scaling o f stability parameters that determine the time complexity of these algorit hms.\n\n Nevertheless\, quantum advantage is not restricted to time comple xity\, and there are many known examples of advantages in terms of other r esources such as communication or space. \n\n We prove that for generic t asks such as inference and gradient computation in distributed\, parameter ized models\, networked quantum computers require exponentially less comm unication compared to any classical analog. This advantage applies to cert ain graph networks\, and we demonstrate that these achieve competitive pe rformance with state of the art neural networks on standard benchmarks. In addition\, we prove that for some distributed computations\, encoding cla ssical data in quantum states prevents it from being re-used for repeated computation. This is a consequence of the destructive nature of quantum me asurement\, and does not require the use of any cryptographic primitives or computational assumptions. This phenomenon could have implications in i ncentivising the creation of datasets and the design of data markets.\n\n \n  \n\n//indico.ictp.it/event/10767/ LOCATION:ICTP Common Area\, ex SISSA building\, second floor URL://indico.ictp.it/event/10767/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar: The leading eigenpair of sparse directed networks DTSTART;VALUE=DATE-TIME:20240808T123000Z DTEND;VALUE=DATE-TIME:20240808T133000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10771@ictp.it DESCRIPTION:\n Understanding the relationship between the structure of spa rse random networks and their spectra is one of the central and long-stand ing focus of network theory. Spectral properties of networks determine the performance of algorithms\, phase transitions\, and the linear stability of complex biological systems\, such as neural networks and ecosystems. In this talk\, I will present simple analytic results for the leading eigenv alue of directed random networks and for the statistics of the correspondi ng leading eigenvector. Specifically\, the leading eigenvalue undergoes a gap-gapless transition as a function of the network parameters\, while the leading eigenvector may become localized. These results provide a clear u nderstanding of the impact of the network structure on the leading eigenpa ir of directed networks\, which is relevant for characterizing the phase d iagram of nonlinear complex systems with sparse interactions. Finally\, I will show that even in highly connected networks\, moderate degree fluctua tions can lead to a breakdown of Gaussian random-matrix universality.\n\n/ /indico.ictp.it/event/10771/ LOCATION:ICTP Common Area\, ex SISSA building\, second floor URL://indico.ictp.it/event/10771/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - A complex systems view on plasmid-bacteria interacti ons DTSTART;VALUE=DATE-TIME:20241016T120000Z DTEND;VALUE=DATE-TIME:20241016T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10809@ictp.it DESCRIPTION:\n Plasmids are infectious agents abundant in bacteria. By imp osing selective pressure and moving genes between bacteria\, they generate diversity\, shape the structure and function of microbial communities\, a nd drive resistance to antibiotics. However\, to understand these effects\ , we need a deeper understanding of the structure of plasmid communities\, of which we know very little. In this talk\, I will first show what the s tructure of a plasmid genetic similarity network (sampled from empirical d ata) can tell us about the potential transmission of antibiotic resistance . Then\, using an agent-based model\, I will demonstrate how the interplay between the structure of plasmid-microbe infection networks influences co mmunity dynamics\, showcasing the emergent dynamics of these complex ecolo gical systems.\n\n//indico.ictp.it/event/10809/ LOCATION:ICTP Central Area\, Second floor\, ex SISSA building URL://indico.ictp.it/event/10809/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - The importance of being ‘discrete’: advances on lattice random walk theories in arbitrary dimensions and lattices DTSTART;VALUE=DATE-TIME:20241028T130000Z DTEND;VALUE=DATE-TIME:20241028T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10803@ictp.it DESCRIPTION:\n Recent analytical developments to describe diffusive proces ses in discrete space and time have opened up opportunities to study inter action processes that are generally hard to get to within a continuous spa ce-time description. I will describe some of these advances focusing on fi rst-passage dynamics in the presence of multiple reactive or partially rea ctive targets and with multiple inert heterogeneities (spatial disorder). I will also show a generalization of the formalism to when the movement st atistics is correlated\, that is for the one-step non-Markov lattice rando m walk\, and its application to a foraging experiment on ants.\n Referenc es\n[1] L. Giuggioli\, Exact spatio-temporal dynamics of confined lattice random walks in arbitrary dimensions: a century after Smoluchowski and Pó lya\, Physical Review X 10(2):021045 (2020).\n[2] S. Sarvaharman and L. Gi uggioli\, Closed-form solutions to the dynamics of confined biased lattice random walks in arbitrary dimensions\, Physical Review E 102(6):062124 (2 020).\n[3] L. Giuggioli and S. Sarvaharman\, Spatio-temporal dynamics of r andom transmission events: from information sharing to epidemic spread\, J ournal of Physics A: Mathematical and Theoretical 55(37):375005 (2022).\n[ 4] T. Kay and L. Giuggioli\, Diffusion through permeable interfaces: Funda mental equations and their application to first-passage and local time sta tistics\, Physical Review Research 4(3):L032039 (2022).\n[5] D. Das and L. Giuggioli\, Discrete space-time resetting model: application to first-pas sage and transmission statistics\, Journal of Physics A: Mathematical and Theoretical 55 (42):424004 (2022).\n[6] D. Das and L. Giuggioli\, Dynamics of lattice random walk within regions composed of different media and int erfaces\, Journal of Statistical Mechanics: Theory and Experiment 1:013201 (2023).\n[7] D. Marris\, S. Sarvaharman and L. Giuggioli\, Exact spatio-t emporal dynamics of lattice random walks in hexagonal and honeycomb domain s\, Physical Review E 107(5):054139 (2023).\n[8] S. Sarvaharman and L. Giu ggioli\, Particle-environment interactions in arbitrary dimensions: a unif ying analytic framework to model diffusion with inert spatial heterogeneit ies\, Physical Review Research 5(4):043281 (2023).\n[9] D. Marris\, L. Giu ggioli\, Persistent and anti-persistent motion in bounded and unbounded sp ace: resolution of the first-passage problem. New Journal of Physics 26:07 3020 (2024).\n[10] L. Giuggioli\, S. Sarvaharman\, D. Das\, D. Marris and T. Kay\, Multi-target search in bounded and heterogeneous environments: a lattice random walk perspective\, preprint to appear in 'Target Search Pro blems' book (2024)\, arXiv preprint arXiv:2311.00464\n \n \n\n//indico.i ctp.it/event/10803/ LOCATION:ICTP Common area\, ex SISSA building\, second floor URL://indico.ictp.it/event/10803/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Deterministic vs. Stochastic Dynamics in Competitive Communities: Insights from Network Structures and Consumer-Resource Model s DTSTART;VALUE=DATE-TIME:20250210T133000Z DTEND;VALUE=DATE-TIME:20250210T143000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10954@ictp.it DESCRIPTION:Abstract:\n \nThe coexistence of competing species presents a challenging theoretical puzzle. Over the years\, various approaches have been proposed to address this issue - some assuming that the dynamics is p rimarily deterministic\, with the observed species representing a stable e quilibrium\, while others suggest that the dynamics is largely stochastic or chaotic\, leading to an unstable community composition. In recent years \,  a few  studies have attempted to integrate both perspectives.\n \nI n this seminar\, several results related to different aspects of these ana lyses will be presented:\n \n1 - It will be shown that in a deterministic system\, the local community exhibits a nested structure when interaction s are symmetric and a hyperuniform structure when interactions are asymmet ric.\n \n2 - Under neutral dynamics with environmental stochasticity\, th e abundance distribution follows a power law and can support either egalit arian or dominant distributions.\n \n3 - An analysis of consumer-resource models reveals the relationships between interaction strength between two species (which reflects niche overlap) and the correlations in their abun dance and growth rate fluctuations.\n\n//indico.ictp.it/event/10954/ LOCATION:ICTP Common area\, Ex SISSA building\, Second floor URL://indico.ictp.it/event/10954/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Nonequilibrium calorimetry - Part 1 - Introduction DTSTART;VALUE=DATE-TIME:20250311T130000Z DTEND;VALUE=DATE-TIME:20250311T134500Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10978@ictp.it DESCRIPTION:\n In the first introductory part\, we start by reviewing so me basic aspects of equilibrium calorimetry.  In particular\, we illustr ate the presence of a Schottky anomaly and the validity of the Nernst Post ulate for the simplest two-level system.  We use that two-level system to open the avenue toward nonequilibrium calorimetry by introducing the two- level switch (agitated qubit). We show how to compute the heat capacity th ere by solving the Poisson equation sourced\n by the excess heat.\n\n//ind ico.ictp.it/event/10978/ LOCATION:ICTP Adriatico Guest House - Kastler Lecture Hall URL://indico.ictp.it/event/10978/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Nonequilibrium calorimetry - Part 2 - Advanced DTSTART;VALUE=DATE-TIME:20250311T141500Z DTEND;VALUE=DATE-TIME:20250311T150000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10979@ictp.it DESCRIPTION:\n In the second part\, we elaborate on more details for de fining and measuring nonequilibrium heat capacities. After the general def inition\, two computational methods are highlighted. We discuss the main f eatures such as the close-to-equilibrium behavior\, the revealing of kinet ic information\, plus the origin of the possible negativity and the validi ty of an extended Third Law.  We illustrate all that with specific exampl es of Markov jump and Markov diffusion processes.\n\n//indico.ictp.it/even t/10979/ LOCATION:ICTP Adriatico Guest House - Kastler Lecture Hall URL://indico.ictp.it/event/10979/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - The generalized Lotka-Volterra model with network st ructure DTSTART;VALUE=DATE-TIME:20250314T123000Z DTEND;VALUE=DATE-TIME:20250314T133000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10976@ictp.it DESCRIPTION:\n In this seminar I will discuss the effect a network structu re has on the generalized Lotka-Volterra model. First I will review the re sults in the high-connectivity limit where the model can be solved with th e help of heterogeneous mean-field theory. And secondly\, I will show new results for the purely cooperative model on networks with finite connectiv ity. Contrary to the high-connectivity regime\, even the case of random re gular graphs (no heterogeneity of degrees) shows non-trivial behavior.\n\n //indico.ictp.it/event/10976/ LOCATION:ICTP Common Area\, Old SISSA building\, second floor URL://indico.ictp.it/event/10976/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Active from active DTSTART;VALUE=DATE-TIME:20250319T130000Z DTEND;VALUE=DATE-TIME:20250319T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-10994@ictp.it DESCRIPTION:Abstract:\n Can activity be transmitted from smaller to larger scales? We show such a transfer from a homogeneous active medium to a New tonian spherical probe. The active medium consists of faster and dilute se lf-propelled particles\, modeled as run-and-tumble particles in 1D or as a ctive Brownian particles in 2D. We derive the reduced fluctuating dynamics of the probe valid for arbitrary probe velocity\, characterized by a nonl inear friction and a velocity-dependent noise. There appear several distin ct regimes: a standard regime where the probe exhibits passive Brownian mo tion\, and peculiar active regimes where the probe becomes self-propelled with high persistence\, and its velocity distribution begets peaks at non- zero values.\n Those peaks and their persistence are quantitatively obtain ed and are perfectly confirmed by numerical simulations.\n  In 1D\, a sof t coupling is necessary\, whereas in 2D\, more realistic interactions\, su ch as Lennard-Jones\, suffice.\n Our study thus reveals how active motion can be transmitted between different scales solely via the induced frictio n and noise.Joint work with Jihui Pei -  KU Leuven\n\n\n//indico.ictp.it/ event/10994/ LOCATION:ICTP Common Area Old SISSA building Second floor URL://indico.ictp.it/event/10994/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS-TLQS Seminar: Biased Transmission Drives an Accuracy-Consensus Tradeoff in Collective Decision-making DTSTART;VALUE=DATE-TIME:20250507T083000Z DTEND;VALUE=DATE-TIME:20250507T093000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11027@ictp.it DESCRIPTION:\n Abstract:Collective decision-making can be more accurate th an individual decision-making\, due to the ability of groups to aggregate information from many individuals. However\, information aggregation is vu lnerable to a variety of social and cognitive biases which can undermine t he "wisdom of the crowds" and result in bad decisions. Here\, we study the effect of two very hard-to-avoid kinds of biases: the tendency of similar individuals to be connected\, and the tendency to share some kinds of inf ormation more than others. We build mathematical models of a society compo sed of rational agents that use Bayesian inference to learn about the worl d\, combining their direct observations with information they obtain from other agents. We include two forms of transmission bias: (1) assortment in communication networks\, where individuals preferentially connect with si milar others\, and (2) selective sharing of signals. We identify condition s under which these effects are sufficient to generate stable divergence o f beliefs across the population--either in the form of polarization or the entire population being misled. We show that when the population is balan ced between types of agents\, the collective is most sensitive to the effe cts of assortment. This is also the condition when the collective is  mos t responsive to the environment. Therefore\, belief polarization is most l ikely under conditions that  support the greatest collective ability to s ense the truth. Our results thus identify tradeoffs between accuracy and c onsensus in collective decision-making and highlight structural limits to collective intelligence\, even in an idealized setting.\n\n//indico.ictp.i t/event/11027/ LOCATION:ICTP Common Area\, old SISSA building\, 2nd floor URL://indico.ictp.it/event/11027/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Flow Of Information In Living Systems DTSTART;VALUE=DATE-TIME:20250515T120000Z DTEND;VALUE=DATE-TIME:20250515T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11039@ictp.it DESCRIPTION:\n Abstract:\n Living entities in a group communicate and tran sfer information to one another for a variety of reasons. It might be for foraging food\, migration\, or escaping from obstacles and threats that m ay appear suddenly. They do so by interacting with each other and also wit h the environment. Statistical mechanics and information theory can be use ful to develop the tools to quantify and analyse the flow of information a mong the living entities. The living entities can be modelled as active (i .e. self-propelling) particles. Here we consider a group of active particl es under confinement. First we will show that the self-organisation of the particles can crucially depend on whether the confinement is soft or hard . Then we quench the trap boundary from soft to hard instantaneously. Con sequently the self-organised cluster of the active particles\, which was s table when the boundary was soft\, becomes unstable and undergoes extreme deformation after the quench to find another stable configuration suitable for the hard boundary. Finally we will quantify the information regardin g the quench that flows throughout the individual particles of the deformi ng cluster and show that the flow spans the whole cluster\, propagating ba llistically.\n  \n\n//indico.ictp.it/event/11039/ LOCATION:ICTP Common Area\, old SISSA building\, second floor URL://indico.ictp.it/event/11039/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Flow of Information in Living Systems DTSTART;VALUE=DATE-TIME:20250515T120000Z DTEND;VALUE=DATE-TIME:20250515T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11040@ictp.it DESCRIPTION:Abstract:\n Living entities in a group communicate and transfe r information to one another for a variety of reasons. It might be for for aging food\, migration\, or escaping from obstacles and threats that may a ppear suddenly. They do so by interacting with each other and also with th e environment. Statistical mechanics and information theory can be useful to develop the tools to quantify and analyse the flow of information among the living entities. The living entities can be modelled as active (i.e. self-propelling) particles. Here we consider a group of active particles u nder confinement. First we will show that the self-organisation of the par ticles can crucially depend on whether the confinement is soft or hard. Th en we quench the trap boundary from soft to hard instantaneously. Conseque ntly the self-organised cluster of the active particles\, which was stable when the boundary was soft\, becomes unstable and undergoes extreme defor mation after the quench to find another stable configuration suitable for the hard boundary. Finally we will quantify the information regarding the quench that flows throughout the individual particles of the deforming clu ster and show that the flow spans the whole cluster\, propagating ballisti cally.\n\n//indico.ictp.it/event/11040/ LOCATION: URL://indico.ictp.it/event/11040/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Connection of feces to Community Ecology. Theoretica l study about the dung beetle population. DTSTART;VALUE=DATE-TIME:20250528T120000Z DTEND;VALUE=DATE-TIME:20250528T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11044@ictp.it DESCRIPTION:\n Abstract:\n Biological invasion is a major problem in the A nthropocene as it causes biodiversity loss. Not only do dung beetles serve as the primary decomposers of animal feces\, but they also play critical roles in numerous ecosystem functions\, so they seemed as key species in t errestrial ecosystem. Invasive mammals have a more significant impact than other taxonomic groups. Non-native mammals introduce their feces into nat ive ecosystems\, as new resources for coprophagous animals. Understanding how non-native mammal invasions affect the dynamics of native dung beetle communities is a critical ecological issue. In this study\, I developed a novel population dynamics model and analyzed it using numerical calculatio ns. In the scenario I explore\, non-native mammals invade a native communi ty where native mammals\, native generalist dung beetles\, and native spec ialist dung beetles coexist. generalist can use invasive mammal feces\, bu t specialist cannot. After the mammal invasion\, specialist populations de clined as invasive mammal carrying capacities increased along with dung be etle competition coefficients. Generalist prefer native feces\, specialist declined more. This study suggests that the possibility of coexistence of the native dung beetle community decreases with non-native mammal invasio n.\n I also speak about field study focusing on deer feces and gut microbi ome. I'd like to introduce you to the world of animals surrounded feces\n\ n//indico.ictp.it/event/11044/ LOCATION:ICTP Common area\, Old SISSA building\, second floor URL://indico.ictp.it/event/11044/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - An algebraic approach to the motion of life DTSTART;VALUE=DATE-TIME:20250605T130000Z DTEND;VALUE=DATE-TIME:20250605T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11046@ictp.it DESCRIPTION:Abstract:\n Motion is a fundamental principle of the living. F lows of energy or biomass serve as the fundamental components of motion\, whose interactions give rise to various events. These flows are not isolat ed\, as their surroundings constantly influence them. Thus\, the Open Syst em Dynamics (OSD) analytic method guides the mathematical modelling of the se phenomena [1]\, [2]. OSD finds a suitable representation through Stocha stic Analysis in both commutative and non-commutative versions. A common a nd simpler approach to these representations is obtained through algebraic probability spaces and stochastic dynamics therein\, which allows to desc ribe basic motion patterns.\n  \nThis lecture is based on a current joint work with Pablo Marquet\, Towards a noncommutative mathematical theory of living systems.\n References[1] Carlo Cellucci. “From Closed to Open S ystems. Philosophy of mathematics”. In: Proceedings of the 15th Internat ional Wittgenstein Symposium\, Wien (Hoeder-Pichler-Tempky). 1993\, pp. 20 6–220.[2] Carlo Cellucci. "Rethinking Knowledge\, The Heuristic View" Sp ringer\, 2017.\n\n//indico.ictp.it/event/11046/ LOCATION:ICTP Common Area\, Old SISSA building\, Second floor URL://indico.ictp.it/event/11046/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Decoding thermosensation and associated response mec hanisms in Caenorhabditis elegans DTSTART;VALUE=DATE-TIME:20250625T090000Z DTEND;VALUE=DATE-TIME:20250625T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11067@ictp.it DESCRIPTION:Abstract:\n All living organisms need to respond to changing e nvironmental signals for their survival. Therefore\, organisms have develo ped elaborate intracellular and intercellular signaling mechanisms to resp ond appropriately to environmental changes [1\, 2]. In this talk\, I will focus on our study of signal-response mechanisms associated with sensing t emperature changes using the roundworm\, C. elegans\, as a model system. I n the first part of the talk\, I will discuss a theoretical model that hel ps in understanding how temperature fluctuations are relayed across the th ermosensory AFD neuron through the cGMP pathway leading to changes in Calc ium response dynamics [3]. In the second part of my talk\, I will discuss our experimental study where we look into how the heat shock response dyna mics is affected by the AFD neuron specific receptor guanylate cyclases (r GCs). Through behavioral assays and molecular biology studies at the mRNA and protein levels\, we find that the rGCs in the neuron play a very impor tant role in upregulating the small heat shock proteins present in the oth er cells during high temperature (>30 deg celcius) conditions [4].\n \n[1 ] Sushmita Pal and Rati Sharma\, Transcription factors and chaperone prote ins play a role in launching a faster response to heat stress and aggregat ion\, Comp. Biol. and Chem. 93\, 107534 (2021)\n \n[2] Ayush Ranawade\, R ati Sharma and Erel Levine\, Lessons Learned from Two Decades of Modeling the Heat-Shock Response\, Biomolecules 12\, 1645 (2022)\n \n[3] Abhilasha Batra and Rati Sharma\, Neuronal Decoding of Temperature Signals in Caeno rhabditis elegans Under review  \nBiorxiv: doi: https://doi.org/10.1101/2 025.01.27.634994\n \n[4] Abhilasha Batra and Rati Sharma\, Guanylyl cycla se signaling in AFD neurons regulates systemic stress resilience in C. ele gans\, Under review\n\n//indico.ictp.it/event/11067/ LOCATION:ICTP Common Area Old SISSA building Second floor URL://indico.ictp.it/event/11067/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Non-Equilibrium Statistical Mechanics of/for AI DTSTART;VALUE=DATE-TIME:20250626T090000Z DTEND;VALUE=DATE-TIME:20250626T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11066@ictp.it DESCRIPTION:Abstract:\n\n \n This talk presents a unifying applied mathema tics/theoretical physics framework that bridges core components of modern generative AI -- diffusion models\, reinforcement learning\, and transform ers -- through the lens of contemporary applied mathematics. Central to th is framework are the concepts of Decision Flows and Path Integral Diffusio ns\, which offer structured approaches to sequential sampling over discret e\, continuous\, and hybrid spaces. These approaches are rooted in Green-f unction-based control\, Schroedinger bridges\, and non-equilibrium statist ical physics.\n Building on recent work\, we explore analytically tractabl e and algorithmically efficient regimes -- often requiring minimal use of neural networks -- where sampling from complex distributions becomes both explainable and extrapolative. We highlight connections between score-base d diffusion\, linearly-solvable Markov Decision Processes\, and energy-bas ed models\, including emerging insights into phase transitions in generati ve AI (e.g.\, memorization and speciation dynamics).\n Applications span i nference/sampling in Ising models\, CIFAR-10 image generation\, physics-in formed reinforcement learning in turbulent flows\, and auto-regressive mod eling of statistical hydrodynamics. We also touch on decision-making under uncertainty in energy systems.\n\n\n//indico.ictp.it/event/11066/ LOCATION:ICTP Common Area Old SISSA building Second floor URL://indico.ictp.it/event/11066/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - An introduction to diffusion models in generative ma chine learning DTSTART;VALUE=DATE-TIME:20250626T080000Z DTEND;VALUE=DATE-TIME:20250626T090000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11068@ictp.it DESCRIPTION:Abstract:\n\n Diffusion models are a class of generative model s in machine learning that iteratively transform data into noise through a forward diffusion process\, typically converging toward a Gaussian distri bution. A corresponding reverse process then reconstructs data-like sample s by denoising these vectors step-by-step. This pair of processes resemble s coarse-graining and fine-graining operations. In this talk\, I will prov ide a brief introduction to the principles and mechanics of diffusion mode ls\, focusing on how they generate realistic samples from noise.\n\n//indi co.ictp.it/event/11068/ LOCATION:ICTP Common Area Old SISSA building Second floor URL://indico.ictp.it/event/11068/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - A Complex Systems Physics Framework for Cortical Sta tes: From E/I Balance to Emergent Oscillations DTSTART;VALUE=DATE-TIME:20250704T090000Z DTEND;VALUE=DATE-TIME:20250704T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11070@ictp.it DESCRIPTION:Abstract:\n Understanding the complex dynamics of excitation a nd inhibition within the cortex offers crucial insights into neural ensemb le behavior across multiple states\, from wakefulness to sleep. By leverag ing tools from dynamical systems and analogies from statistical physics\, this body of work explores the spatiotemporal balance between excitation a nd inhibition\, highlighting how these dynamics contribute to both local a nd global cortical activity patterns. The observed symmetry breaking at cr itical moments\, reminiscent of phase transitions in physical systems\, un derscores the sensitivity of cortical networks to small perturbations\, wh ich can cascade into larger-scale oscillations and state changes.\n High-f requency oscillations\, including beta and gamma rhythms\, emerge from thi s intricate interplay and propagate systematically across the cortex. Thes e oscillations are especially prominent during slow-wave sleep\, where the y likely play a role in cognitive functions such as memory consolidation. By integrating ensemble analyses across different behavioral states—wake fulness\, sleep\, and transitions—this research illuminates how excitati on and inhibition orchestrate the neocortex’s ability to dynamically bal ance stability and flexibility\, driving fundamental processes such as osc illatory rhythms and coordinated network activity.\n This work not only ad vances our understanding of the neural processes underlying normal cortica l function but also offers insights into pathological conditions\, such as seizures\, where this balance is disrupted. These findings bridge the gap between neural biophysics and large-scale functional dynamics\, offering a fresh perspective on how cortical states self-organize through emergent\ , multiscale processes.\n References:\n1. Symmetry’s Edge in Cortical D ynamics: Multiscale Dynamics of Ensemble Excitation and Inhibition (https: //arxiv.org/html/2306.11965v4) \n2. Ensemble Inhibition and Excitation in the Human Cortex: An Ising-model Analysis with Uncertainties (https://jou rnals.aps.org/pre/abstract/10.1103/PhysRevE.99.032408) \n3. Dynamic Balan ce of Excitation and Inhibition in Human and Monkey Neocortex (https://www .nature.com/articles/srep23176) \n4. High-frequency Oscillations in Human and Monkey Neocortex During the Wake–Sleep Cycle (https://www.pnas.org/ doi/abs/10.1073/pnas.1523583113) \n5. Spatiotemporal Dynamics of Neocorti cal Excitation and Inhibition During Human Sleep (https://www.pnas.org/doi /full/10.1073/pnas.1109895109)\n\n//indico.ictp.it/event/11070/ LOCATION:ICTP Common area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/11070/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Dynamics and functions of the p53 signaling network in response to various stresses DTSTART;VALUE=DATE-TIME:20250801T090000Z DTEND;VALUE=DATE-TIME:20250801T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11092@ictp.it DESCRIPTION:Abstract:\n As one of the most critical tumor suppressors\, th e p53 protein orchestrates cellular physiological activities through an in tricate signaling network. Crucially\, the temporal evolution of p53 conce ntration significantly regulates cellular stress responses. Through theore tical modeling and numerical simulations\, we systematically investigated the dynamics and functions of the p53 network in response to diverse signa ls\, such as DNA damage\, oncogene activation\, glucose starvation\, and h ypoxia. Here\, I will present our representative research\, highlighting t he critical link between p53 dynamics and function.\n \nReferences：\n[1 ]  X.-P. Zhang\, F. Liu and W. Wang. Two-phase dynamics of p53 in the DNA damage response. Proc. Natl. Acad. Sci. USA 108\, 8990 (2011).\n[2]  R.Y ang\, B. Huang\, Y. Zhu\, Y. Li\, F. Liu and J. Shi. Cell type-dependent b imodal p53 activation engenders a dynamic mechanism of chemoresistance. Sc i. Adv. 4\, eaat5077 (2018).\n[3]  Y. Zhou and F. Liu.  Coordination of the AMPK\, Akt\, mTOR\, and p53 pathways under glucose starvation. Int. J. Mol. Sci. 23\, 14945 (2022).\n[4]  P. Wang\, X.-P. Zhang\, F. Liu and W. Wang. Progressive deactivation of hydroxylases controls hypoxia-inducible factor-1α-coordinated cellular adaptation to graded hypoxia. Research 8\ , 0651 (2025).\n\n//indico.ictp.it/event/11092/ LOCATION:ICTP Common Area Old SISSA building Second floor URL://indico.ictp.it/event/11092/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Spatio-temporal Selection Dynamics: From Evolutionar y Graphs to Reaction–Diffusion Equations DTSTART;VALUE=DATE-TIME:20250806T090000Z DTEND;VALUE=DATE-TIME:20250806T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11093@ictp.it DESCRIPTION:\n I will discuss stochastic models of evolutionary dynamics i n structured populations\, focusing on evolutionary graph theory and its e xtension to game-theoretic interactions. I will present recent work showin g how spatial Moran processes converge\, in the large-population limit\, t o continuum models that extend or deviate from the classical Fisher–Kolm ogorov equation. In certain regimes\, these limits yield reaction–diffus ion dynamics related to the KPZ universality class. I will also touch on e xtensions to evolutionary games\, evolution in heterogeneous media and app lications to spatial modelling of cancer therapies\, including glioblastom a.\n\n//indico.ictp.it/event/11093/ LOCATION:ICTP Common Area Old SISSA building Second floor URL://indico.ictp.it/event/11093/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Modeling healthy and pathological aging DTSTART;VALUE=DATE-TIME:20250827T083000Z DTEND;VALUE=DATE-TIME:20250827T093000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11101@ictp.it DESCRIPTION:As we age\, our brains undergo substantial structural and func tional changes\, which can be either accelerated or slowed by various fact ors. In this talk\, I will demonstrate how these processes can be modeled using non-invasive neuroimaging data. I will present results obtained thro ugh machine learning to calibrate brain-age estimation models\, which can then be used to investigate how different diseases and environmental facto rs disrupt healthy aging trajectories. Finally\, I will discuss how biophy sically grounded models based on coupled dynamical systems can shed light on the mechanisms underlying these accelerated aging processes.\n\n//indic o.ictp.it/event/11101/ LOCATION:ICTP Common Area Old SISSA building 2nd floor URL://indico.ictp.it/event/11101/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - The Global Phenospace of Vascular Plants DTSTART;VALUE=DATE-TIME:20250915T093000Z DTEND;VALUE=DATE-TIME:20250915T103000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11110@ictp.it DESCRIPTION:This is an informal talk that will present the idea of the glo bal phenospace\, focusing on how this concept is defined in terms of quant itative data and on open questions and challenges.\n\n//indico.ictp.it/eve nt/11110/ LOCATION:ICTP Common Area\, Old SISSA building\, 2nd floor URL://indico.ictp.it/event/11110/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Controlled and Uncontrolled Growth: Mathematical Mod els of Regeneration and Tumor Protection DTSTART;VALUE=DATE-TIME:20251001T090000Z DTEND;VALUE=DATE-TIME:20251001T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11189@ictp.it DESCRIPTION:\n Tissues in higher organisms are organized as hierarchically structured cell populations\, where only a small number of cells\, the st em cells\, have the capacity for infinite self-renewal. In this talk\, we explore how stochastic models are used to show that such hierarchies prote ct against uncontrolled\, cancer-like growth.\n  \n However\, this design that protects most tissues has a drawback\, slow regeneration\, leading t o an exception: the liver. After an injury\, fully differentiated hepatocy tes can re-enter the cell cycle and proliferate without requiring stem cel ls or progenitors to produce new cells.\n  \n To explore how this mechani sm is controlled\, we use single-cell RNA-seq to dissect the early events of liver regeneration. We show how macrophage signaling to hepatocytes dri ves the process\, and how mathematical models\, calibrated on the single-c ell data\, allow in-silico knockouts that identify a key regenerative sign al.\n  \n Finally\, we explore how these two concepts\, restricted self-r enewal for cancer protection and direct hepatocyte division for rapid repa ir\, can be reconciled.\n  \n  \n\n//indico.ictp.it/event/11189/ LOCATION:ICTP Common Area Old SISSA building Second floor URL://indico.ictp.it/event/11189/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - A Statistical Physics View of Species Vulnerability in Tropical Forests and Beyond DTSTART;VALUE=DATE-TIME:20251006T120000Z DTEND;VALUE=DATE-TIME:20251006T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11185@ictp.it DESCRIPTION:\n Understanding why some species persist while others vanish remains a central challenge in ecology. Traditional approaches either rely on mechanistic models that are conceptually clear but hard to apply to re al-world scenarios\, or on statistical tools that lack generality and inte rpretability. Here\, we introduce a framework rooted in the statistical ph ysics of complex systems but designed for ecological applicability. At its core is a single measurable quantity - the competitive balance - that syn thesizes the processes shaping persistence: dispersal diversity\, interspe cific competition\, and abundance. Without requiring direct measurement of species traits or detailed dispersal parameters\, the metric captures how the spatial structure and heterogeneity of communities translate into spe cies-level vulnerability. Our analyses reveal a fundamental trade-off: for a given rarity\, greater heterogeneity in dispersal strategies reduces vu lnerability. We validate the framework with tropical forest data\, but its scope extends across ecosystems. By linking statistical physics with func tional perspectives on biodiversity\, this approach offers an interpretabl e and scalable tool to assess species vulnerability in the Anthropocene.\n  \n\n//indico.ictp.it/event/11185/ LOCATION:ICTP Central Area\, second floor\, old SISSA building URL://indico.ictp.it/event/11185/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Limits of Inference in Stochastic Models for Comple x Systems DTSTART;VALUE=DATE-TIME:20251021T120000Z DTEND;VALUE=DATE-TIME:20251021T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11204@ictp.it DESCRIPTION:\n Simple stochastic models are widely used to describe comple x phenomena\, as they provide an interpretable connection between minimal microscopic descriptions and macroscopic statistical patterns. A central c hallenge in this approach is selecting the appropriate model—and its par ameters—to accurately capture specific datasets. In this talk\, I presen t a framework based on change-of-measure theory and bridge processes to pe rform parametric inference and assess model distinguishability using path statistics. With this tool\, we identify optimal sampling protocols that m inimize errors in parameter estimation\, while also revealing fundamental limits of inference: the ability to discriminate between competing models is intrinsically constrained by data quality (e.g.\, sampling frequency an d dataset size). We illustrate the framework with applications to diverse datasets (particle trapped with optical tweezers\, human microbiome\, top ic mentions in social media\, and forest population dynamics) showing how inference limits manifest in practice and inform ongoing modeling debates . The talk aims to give an overview of the state of the art in model infer ence\, highlighting both current methodologies and their inherent limitati ons.\n  \n  \n\n//indico.ictp.it/event/11204/ LOCATION:ICTP Common Area\, Old SISSA building Second floor URL://indico.ictp.it/event/11204/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Limits of Inference in Stochastic Models for Comple x Systems DTSTART;VALUE=DATE-TIME:20251021T120000Z DTEND;VALUE=DATE-TIME:20251021T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11211@ictp.it DESCRIPTION:\n Simple stochastic models are widely used to describe comple x phenomena\, as they provide an interpretable connection between minimal microscopic descriptions and macroscopic statistical patterns. A central c hallenge in this approach is selecting the appropriate model—and its par ameters—to accurately capture specific datasets. In this talk\, I presen t a framework based on change-of-measure theory and bridge processes to pe rform parametric inference and assess model distinguishability using path statistics. With this tool\, we identify optimal sampling protocols that m inimize errors in parameter estimation\, while also revealing fundamental limits of inference: the ability to discriminate between competing models is intrinsically constrained by data quality (e.g.\, sampling frequency an d dataset size). We illustrate the framework with applications to diverse datasets (particle trapped with optical tweezers\, human microbiome\, top ic mentions in social media\, and forest population dynamics) showing how inference limits manifest in practice and inform ongoing modeling debates . The talk aims to give an overview of the state of the art in model infer ence\, highlighting both current methodologies and their inherent limitati ons.\n\n//indico.ictp.it/event/11211/ LOCATION:ICTP Common Area\, Old SISSA building Second floor URL://indico.ictp.it/event/11211/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Controlling collective phases via Subsystem Resettin g in many-body interacting systems DTSTART;VALUE=DATE-TIME:20251022T090000Z DTEND;VALUE=DATE-TIME:20251022T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11201@ictp.it DESCRIPTION:\n For practical applications\, one is often interested in int eracting systems in a phase which is thermodynamically unstable in parame ter regimes of interest. In traditional approaches\, such phases are stabi lised through intervening in the dynamics of all system constituents or i ntroducing additional interactions. In this work\, we present an alternat ive approach for stabilising and systematically manipulating phases in the context of many-body interacting systems\, namely\, subsystem resetting.  The idea is simple: only a chosen subpart of the full system is stochast ically reset to a fixed configuration\, while the rest of the system evolv es freely\, and we ask what happens to a specific order parameter at stea dy state. This avoids global interventions and fine-tuning of couplings.  We have applied this protocol across a wide variety of models\, namely\, m ean-field spin models and the noisy Kuramoto oscillator system. In this ta lk\, after discussing the protocol\, I will focus on the application to K uramoto oscillators and one of the spin models. I will demonstrate how r esetting a subpart can have robust control of the phase diagram of the ori ginal model for the rest of the system\, by adjusting the reset frequency\ , subsystem size\, and reset configuration. These results open up a versa tile framework for stabilising or suppressing targeted phases in complex i nteracting systems.\n  \n\n//indico.ictp.it/event/11201/ LOCATION:ICTP Common Area\, Old SISSA building URL://indico.ictp.it/event/11201/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS SEMINAR - Decoding behavior: optimization and inference of dec ision-making processes DTSTART;VALUE=DATE-TIME:20251028T130000Z DTEND;VALUE=DATE-TIME:20251028T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11214@ictp.it DESCRIPTION:\n I will analyze some selected examples of complex behaviors through the lens of decision-making theory. The aim is to identify minimal models that explain how agents process information and perform actions (w ith Giorgio Nicoletti).\n\n//indico.ictp.it/event/11214/ LOCATION:ICTP Common Area\, Old SISSA building\, second floor URL://indico.ictp.it/event/11214/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Statistical physics of deep learning: Optimal learni ng of a multi-layer perceptron near interpolation DTSTART;VALUE=DATE-TIME:20251107T100000Z DTEND;VALUE=DATE-TIME:20251107T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11217@ictp.it DESCRIPTION:\n For three decades\, statistical physics has framed neural-n etwork analysis\, but its reach to expressive\, feature-learning deep mode ls was unclear. We answer yes by studying supervised learning in fully con nected multi-layer nets whose hidden layers scale with input dimension—f avoring feature learning over ultra-wide kernels while remaining more expr essive than narrow or fixed-weight models—in the challenging interpolati on regime where parameters and data are comparable. Using a matched teache r–student setup\, we characterize fundamental performance limits and the sufficient statistics learned as data grows. The analysis uncovers rich p henomenology with multiple learning transitions: with enough data\, optima l performance requires “specialization” of the student to the target\, yet practical training can be trapped in sub-optimal solutions. Specializ ation is inhomogeneous—spreading from shallow to deep layers and unevenl y across neurons—and deeper targets are intrinsically harder. Though der ived in a Bayesian-optimal setting\, the insights on nonlinearity\, depth\ , and finite (proportional) width likely generalize.\n  \n  \n  \n\n//i ndico.ictp.it/event/11217/ LOCATION:ICTP Common area\, second floor\, old SISSA building URL://indico.ictp.it/event/11217/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Evolution of anticipation under partially predictab le environments DTSTART;VALUE=DATE-TIME:20251119T130000Z DTEND;VALUE=DATE-TIME:20251119T140000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11231@ictp.it DESCRIPTION:\n Microbes in fluctuating environments must balance rapid gro wth against fast adaptation (short lag times). Using proteome allocation m odels\, we show evolution drives populations to an evolutionary stable str ategy that optimally balances this trade-off. In predictable environments\ , populations evolve to "learn" statistical patterns by pre-allocating the ir proteome according to environmental transition probabilities. This demo nstrates how microbes use the proteome as a memory to encode and exploit e nvironmental patterns without neural computation.\n\n//indico.ictp.it/even t/11231/ LOCATION:ICTP Common Area\, Second floor\, old SISSA building URL://indico.ictp.it/event/11231/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Cooperative mixing through hydrodynamic interactions in Stylonychia lemnae DTSTART;VALUE=DATE-TIME:20251210T100000Z DTEND;VALUE=DATE-TIME:20251210T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11239@ictp.it DESCRIPTION:\n How do aquatic microorganisms optimize feeding: is it more effective to swim through food\, remain anchored while pumping food toward themselves\, or join others to form a feeding colony? Despite how widespr ead and fundamental these strategies are\, we still lack a systematic unde rstanding of which is most efficient under different environmental conditi ons. This gap largely stems from the scarcity of behavioral studies combin ed with experimental investigations of how microorganism-generated flows s tir the surrounding fluid and\, in doing so\, reshape nearby resource patc hes.\n In this study\, we focus on Stylonychia lemnae and found\, first\, that these ciliates employ chemotaxis to cluster around food patches\, yet still remaining physically independent. Then\, as many cells cluster on t he food patch\, their individual feeding flows interact\, generating a cha otic collective flow throughout the cluster. Using a combination of experi ments and simulations\, we measured and predicted the full active chaotic flow. We found that it efficiently stirs and disperses food particles thro ughout the entire cluster\, thereby increasing the effective patch size an d helping still-foraging cells to locate food more quickly. This form of c ooperative behavior highlights the surprising complexity of feeding strate gies at the microscale.\n\n \n  \n\n//indico.ictp.it/event/11239/ LOCATION:ICTP Common area\, Old SISSA building\, Second floor URL://indico.ictp.it/event/11239/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Memory Capacity\, Novelty Encoding\, and Continual L earning in Biologically Plausible Neural Networks DTSTART;VALUE=DATE-TIME:20260213T100000Z DTEND;VALUE=DATE-TIME:20260213T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11260@ictp.it DESCRIPTION:\n I will present two projects and briefly discuss my interest in continual learning.\n  \n The first project is a theoretical study of memory capacity in heterogeneous neural networks. We generalize classical Hopfield-model results to networks with heterogeneous neuron coding level s and arbitrary connectivity\, deriving an analytical expression for maxim al memory capacity. Although heterogeneity typically reduces capacity\, we show that capacity can be preserved when coding levels and in-degree are positively correlated. This principle holds across multiple biologically r elevant settings and can be experimentally tested. In addition\, in a two- layer network representing the hippocampal CA3–DG circuit\, we find that a quasi-indexing encoding scheme in the DG maximizes capacity\, extending the hippocampal index theory of memory.\n  \n The second project is an e xperimental study of how the basal forebrain encodes salience and novelty. We identify a population of basal forebrain neurons that exhibit phasic b ursting responses to surprising external events\, including unexpected rew ards\, novel stimuli\, and sequence violations. These neurons do not encod e reward prediction errors\, but instead signal surprise and salience\, su ggesting a mechanism by which the basal forebrain primes cortical circuits for enhanced processing of important information.\n\n//indico.ictp.it/eve nt/11260/ LOCATION:ICTP Common Area\, Old SISSA building Second floor URL://indico.ictp.it/event/11260/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Hierarchical inference in the brain as a chain of 7 “neurons” (excitable integrators): criticality\, activity and predict ion errors\, power-laws\, and Up-Down states DTSTART;VALUE=DATE-TIME:20260318T100000Z DTEND;VALUE=DATE-TIME:20260318T110000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11313@ictp.it DESCRIPTION:\n We describe a chain of unidirectionally coupled adaptive ex citable elements slowly driven by a stochastic process from one end and o pen at the other end\, as a minimal toy model of unresolved irreducible u ncertainty in a system performing inference through a hierarchical model. Threshold potentials adapt slowly to ensure sensitivity without being was teful. Activity and energy are released as intermittent avalanches of puls es with a distribution largely independent of the exogenous input form.  We find that subthreshold bistability closely resembles empirical measurem ents of intracellular membrane potential\, and suggest that critical corti cal cascades emerge from a trade-off between metabolic power consumption a nd performance requirements in a critical world\, and that the temporal s caling patterns of brain electrophysiological recordings ensue from weight ed linear combinations of subthreshold activities and pulses from differen t hierarchy levels.\n\n//indico.ictp.it/event/11313/ LOCATION:ICTP Common Area\, Old SISSA building\, second floor URL://indico.ictp.it/event/11313/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Model of Sperm–Oocyte Penetration Mechanisms DTSTART;VALUE=DATE-TIME:20260515T113000Z DTEND;VALUE=DATE-TIME:20260515T123000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11352@ictp.it DESCRIPTION:\n Existing theories of sperm–oocyte penetration can be clas sified into three main categories: biochemical receptor–mediated mechani sms\, maximum normal stress–based penetration\, and oscillatory relaxati on mechanisms. Previous mechanical probing studies of zona pellucida (ZP) demonstrate increased structural resistance under higher perpendicular loa ding\, implying that penetration is favored along directions of reduced me chanical resistance. Experimental observations indicate that sperm penetra tion of the ZP follows an oblique trajectory. Motivated by these findings\ , a modified mechanical model of ZP deformation is proposed to quantify no rmal and shear stresses generated during plastic sperm–oocyte collision. Numerical simulations were conducted for three distinct sperm distributio n scenarios over the ZP surface. The results show time-dependent oscillati ons of both normal and shear stresses\, with normal stress consistently ex ceeding shear stress across all scenarios. The observed stress fluctuation s may contribute to oscillatory states of the ZP and could play a signific ant role in the mechanical mechanism of sperm penetration. The theory of p hase transition of ZP as a potential mechanism of sperm penetration will b e discussed.\n\n//indico.ictp.it/event/11352/ LOCATION:ICTP Common Area Old SISSA building\, Second floor URL://indico.ictp.it/event/11352/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Cell signaling\, collective rhythms and patterns in embryonic development DTSTART;VALUE=DATE-TIME:20260616T080000Z DTEND;VALUE=DATE-TIME:20260616T090000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11379@ictp.it DESCRIPTION:\n Cells generate\, process and share information to orchestra te the patterning of tissues and organs during embryonic development. An i nteresting example is vertebrate segmentation\, where the axis going from head to tail is subdivided into regular segments that will later form the vertebrae and other tissues. This segmentation is rhythmic\, and it is tho ught that a gene regulatory network is responsible for the rhythm at the c ellular level. At a local level\, intercellular signaling communicates cel ls and synchronizes individual oscillators into a collective oscillation. This collective oscillation produces gene expression waves that traverse t he unsegmented tissue and give rise to segments. In this talk\, I will dis cuss some findings about how this collective rhythm is organized\, from an interdisciplinary perspective that brings together theory and experiment. \n  \n\n//indico.ictp.it/event/11379/ LOCATION:ICTP Common Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/11379/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Lecture Series - From Molecular Motors to Collective Phenomena in Active Matter DTSTART;VALUE=DATE-TIME:20260617T090000Z DTEND;VALUE=DATE-TIME:20260701T103000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11377@ictp.it DESCRIPTION:Professor Mohammad Reza Ejtehadi (Sharif U. of Technology and QLS Associate) will give a series of  three one-hour lectures\, ranging from molecular motors to collective phenomena in the active matter.Lecture 1 - Wed. 17 June 11h00-12h00Molecular Machines/Motors\n- An introduction to active particles and active matter in biology and physics\n- Machines a nd Motors\n- Chemical reactions\n- Catalysts and Enzymes\n- Macromolecules (proteins)\n- Chemically driven Molecular MachinesLecture 2 - Wed. 24 Jun e 11h00-12h00Active Particles\n- Swimmers at Low Reynolds Numbers\n- Chemo taxis\n- Active Brownian Particles\n- Run and TumbleLecture 3 - Wed. 1 Jul y 11h00-12h00Active Matter\n- Interactions between Active Particles\n- Col lective behaviors in Active Matter\n- Ordered and Disordered Phases\n- Hyd rodynamics of active matter\n- Active Polymers\n\n//indico.ictp.it/event/1 1377/ LOCATION:ICTP Central area\, old SISSA building\, 2nd floor URL://indico.ictp.it/event/11377/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Stochastic excursions: from counting observables to inference DTSTART;VALUE=DATE-TIME:20260618T080000Z DTEND;VALUE=DATE-TIME:20260618T090000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11380@ictp.it DESCRIPTION:\n Understanding fluctuations of nonequilibrium observables ac ross stochastic trajectories is essential for various fields of research\, from quantum thermal machines to biological motors. We introduce a framew ork to analyze the statistics of counting observables over the recently in troduced concept of stochastic excursions. Our approach offers analytical expressions for currents\, heat\, entropy production\, and more\, capturin g their correlations and finite-time fluctuations. As main results\, we sh ow how distinct noise sources contribute to fluctuations\, and prove a flu ctuation theorem for excursions that reveals a precision-dissipation trade -off. We illustrate how these results provide new insights for the experim entally realized absorption refrigerator. Lastly\, we show how observing e xcursions can be used to infer thermodynamic quantities\, discussing their connection to previously established methods.\n\n//indico.ictp.it/event/1 1380/ LOCATION:ICTP Common Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/11380/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Fixation Probability in Complex Ecological Landscape s: When Does Heterogeneity Help? DTSTART;VALUE=DATE-TIME:20260625T090000Z DTEND;VALUE=DATE-TIME:20260625T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11384@ictp.it DESCRIPTION:\n A central question in theoretical population biology is whe ther spatial ecological heterogeneity helps or hinders the evolutionary su ccess of mutants. Here I present a unified framework for selection on latt ice graphs with arbitrary environmental heterogeneity\, spanning two regim es: constant selection and frequency-dependent evolutionary games. In the first part\, I calculate the fixation probability of advantageous mutants. I observe that the genotype specificity of the environment determines whe ther heterogeneity amplifies or suppresses selection\, while spatial corre lation in environmental states acts as a secondary determinant of its magn itude. In the second part\, I extend the framework to the evolution of coo peration under spatially varying payoffs\, where spatial correlation becom es the primary determinant\, predicting both the fixation probability of c ooperators and the timescales of selection. I discuss applications to the evolution of drug resistance\, cooperation in microbial communities\, and somatic evolution in epithelial tissues.\n\n//indico.ictp.it/event/11384/ LOCATION:ICTP Common Area\, 2nd floor\, Old SISSA building URL://indico.ictp.it/event/11384/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Maximin Stability in the Graph Model for Conflict Re solution: Forward and Preference Adjustment Approaches DTSTART;VALUE=DATE-TIME:20260702T120000Z DTEND;VALUE=DATE-TIME:20260702T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11385@ictp.it DESCRIPTION:\n Conflicts are inherent to many human interactions and subst antial costs are associated with unresolved conflicts. In this talk\, I wi ll give an introduction about the Graph Model for Conflict Resolution (GMC R)\, which has been successfully applied to model and analyze many real wo rld strategic conflicts. After recalling the main stability concepts used within GMCR\, I will focus on the Maximin stability\, which has been shown to generalize to higher horizons of analysis some classic stability notio ns. I will present the Maximin both in the traditional forward approach th at determines which conflict scenarios are stable given the actions than c an be taken by decision makers and their preferences\, and also in the pre ference adjustment approach\, where a linear integer optimization model to determine minimal preference adjustment to make a desired scenario a Maxi min equilibrium is proposed.\n\n//indico.ictp.it/event/11385/ LOCATION:ICTP Central Area\, 2nd floor\, old SISSA building URL://indico.ictp.it/event/11385/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Flocking in One-dimensional Systems of Interacting Brownian Particles: Synchronous vs Asynchronous Updates DTSTART;VALUE=DATE-TIME:20260703T090000Z DTEND;VALUE=DATE-TIME:20260703T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11388@ictp.it DESCRIPTION:\n We study a one-dimensional system of Brownian particles tha t interact among themselves through local velocity-lignment forces that ar e discrete in time and do not affect the particle speed. The local veloci ty-alignment acts periodically over the system with time period t_a and pr obability P= t_a/t_0\, with t_0 being a control parameter. Depending on th e relation between t_a and t_0\, the velocity-alignment can affect all of the particles synchronously in time or just a few asynchronously. In any case\, as the velocity-alignment events become more frequent\, the system is driven from stationary states close to thermal equilibrium\, to far-fro m-equilibrium ones where the system exhibits spontaneous symmetry breaking and self-organization. Therefore\, our results show that self-propulsion  is not a necessary ingredient to produce the flocking transition even in one-dimensional systems. Above the critical point\, for the synchronous c ase\, the velocity distribution of the particles exhibits periodic oscilla tions in time with period t_a\, and their density distribution shows a bre athing-like behavior. The effects of these phenomena produce a regular sp iking activity in the order parameter as the system approaches to the ther modynamic limit.\n\n//indico.ictp.it/event/11388/ LOCATION:ICTP Common Area\, Second Floor\, old SISSA building URL://indico.ictp.it/event/11388/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Function-specific epistasis shapes evolutionary tra jectories towards antibiotic resistance DTSTART;VALUE=DATE-TIME:20260707T080000Z DTEND;VALUE=DATE-TIME:20260707T090000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11395@ictp.it DESCRIPTION:\n Antibiotic resistance poses a major threat to public health : as pathogens become resistant to known antibiotics\, rendering current t reatments ineffective\, the number of new antibiotics approved for use is decreasing. The consequent need of novel treatment strategies claims for a better understanding of how antibiotic resistance emerges from the evolut ionary processes in bacterial populations\, and the factors that influence the evolution of resistance.\n\n Bacterial populations can rapidly evolve resistance through spontaneous mutations under antibiotic section. Import antly\, pre-existing mutations can influence evolution by constraining or opening evolutionary pathways through epistatic interactions. In some case s\, global epistasis patterns allow evolutionary pathways to be predicted. In other cases\, idiosyncratic epistasis makes evolution less predictable . To systematically study this phenomenon in the context of antibiotic res istance\, we evolved 258 Escherichia coli gene-deletion strains – mimick ing 258 pre-existing mutations – under three antibiotics: trimethoprim\, mecillinam\, and nitrofurantoin. These experiments were performed on a ne w high-throughput platform capable of running 864 parallel automated evolu tion experiments with tight feedback control of population size and select ion pressure\, allowing a quantitative analysis of resistance trajectories . We show that the evolution of antibiotic resistance is highly repeatable \, following a common path across most genetic backgrounds. However\, a mi nority of pre-existing mutations lead to evolutionary trajectories that si gnificantly deviate from this common path. Rather than being predictable f rom global epistasis\, these deviations are modulated by function-specific epistasis: perturbations to specific cellular functions lead to novel evo lutionary trajectories towards resistance. Importantly\, function-specific epistasis often slows down resistance evolution. These findings advance o ur understanding of the molecular mechanisms of resistance evolution in ba cteria and suggest that function-specific epistasis can be exploited as a strategy to combat resistance.\n\n//indico.ictp.it/event/11395/ LOCATION:ICTP Common Area Old SISSA building Second floor URL://indico.ictp.it/event/11395/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Actin Force-Generation in Synthetic Biological Mater ials: Active Droplets and Beating Tails DTSTART;VALUE=DATE-TIME:20260714T080000Z DTEND;VALUE=DATE-TIME:20260714T090000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11394@ictp.it DESCRIPTION:\n Actin is a vital cytoskeletal protein that functions as a m olecular polymerization motor. Within living cells\, actin filaments gener ate mechanical forces (without any molecular motors) through a dynamic bal ance of polymerization and depolymerization in a process known as treadmil ling.\n This machinery allows cells to crawl and reshape themselves. Patho gens like the bacteria Listeria monocytogenes also exploit this mechanism\ , hijacking the host cell’s actin to build propulsion tail networks that drive them forward. Our works reconstruct this treadmilling mechanism wit hin synthetic biological materials. By transitioning from complex cellular environments to controlled synthetic systems\, we explore how biomimetic force generation leads to emergent non-equilibrium phenomena.\n  \n In th is talk\, I will present two complementary approaches that exploit actin t readmilling within synthetic biological materials. First\, I will discuss how local actin filament growth alters the phase-separation dynamics of DN A-nanostar condensates\, driving emergent structural behaviors from drople t reorganization to fragmentation. Second\, I will show how leveraging thi s same motor mechanism allows us to engineer microscopic "beating tails" d riven purely by polymerization. These findings collectively demonstrate ho w simple\, localized molecular mechanisms can be leveraged to drive comple x macroscopic behaviors\, offering a new playground for exploring emergent active matter physics.\n  \n\n//indico.ictp.it/event/11394/ LOCATION:ICTP Common Area Old SISSA building Second floor URL://indico.ictp.it/event/11394/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Physics-Informed Machine Learning for Biological Sys tems DTSTART;VALUE=DATE-TIME:20260715T120000Z DTEND;VALUE=DATE-TIME:20260715T130000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11403@ictp.it DESCRIPTION:\n Physics-informed machine learning (PIML) is emerging as a p owerful framework for modeling complex biological systems by integrating m echanistic knowledge with experimental data. Unlike purely data-driven app roaches\, PIML incorporates differential equations describing biological p rocesses while enabling unknown mechanisms to be learned directly from spa rse\, noisy\, and partially observed measurements. This capability is part icularly valuable in biology\, where mechanistic models are often incomple te and experimental data are limited. In this talk\, I will present recent advances from our group in developing interpretable physics-informed lear ning methods for biomedical applications. I will introduce Universal Physi cs-Informed Neural Networks (UPINNs)\, which extend conventional PINNs by simultaneously estimating unknown parameters and discovering previously un known biological functions from data. I will demonstrate applications in q uantitative systems pharmacology\, drug development\, and cancer biology\, and discuss how these approaches can accelerate mechanistic discovery\, i mprove predictive modeling\, and support the development of next-generatio n biological digital twins.\n\n//indico.ictp.it/event/11403/ LOCATION:ICTP Leonardo Building - Luigi Stasi Seminar Room URL://indico.ictp.it/event/11403/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - Learning beyond lookup tables DTSTART;VALUE=DATE-TIME:20260716T090000Z DTEND;VALUE=DATE-TIME:20260716T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11404@ictp.it DESCRIPTION:\n Decades of research in the laboratory setting has led to bi g advances in our understanding of how animals learn\, and the neural circ uits involved. Surprisingly\, most of such research involves behavioral ta sks that essentially test memorization\, using the same stimuli\, task rul es or motor outputs in the testing phase as those used in the training. Th is contrasts machine learning systems\, where generalization is considered a fundamental test of learning\, and emphasizes performance evaluation wi th new in-distribution or out-of-distribution examples. We are developing behavioral tasks in mice to understand the brain basis of generalization. I will present a concrete example on olfactory learning in mice that invol ves detection of relevant odorants embedded within complex mixtures of dis tractors.\n These experiments suggest that mice have an inductive bias tow ards generalization.\n  \n\n//indico.ictp.it/event/11404/ LOCATION:ICTP Common Area\, old SISSA building\, second floor URL://indico.ictp.it/event/11404/ END:VEVENT BEGIN:VEVENT SUMMARY:QLS Seminar - A primer in Gravity from Entropy DTSTART;VALUE=DATE-TIME:20260917T090000Z DTEND;VALUE=DATE-TIME:20260917T100000Z DTSTAMP;VALUE=DATE-TIME:20260910T124325Z UID:indico-event-11513@ictp.it DESCRIPTION:\n Gravity from Entropy (GfE) is a recently proposed descripti on of gravity arising from a statistical mechanics and information theory action (the GfE action).\n According to this approach gravity is the infor mation captured by the interplay between geometry and matter fields.\n The GfE equations reduce to Einstein equation in the low energy limit\, while the describe modified gravity for high energy and large curvatures.\n In this talk we will provide a gentle introduction to the topic\, focusing on the statistical mechanics aspects of the theory\, and introducing gently the necessary differential geometry elements.\n The presentation is likely to be interactive providing a perspective on the black hole area law and the role of the second principle of thermodynamics in cosmology.\n  \n\n/ /indico.ictp.it/event/11513/ LOCATION:ICTP Common Area\, second floor\, old SISSA building URL://indico.ictp.it/event/11513/ END:VEVENT END:VCALENDAR