Description
Overview
 
This workshop is the second edition of a series of yearly meetings of the Dark Machines community (http://darkmachines.org/), a research collective of physicists and data scientists. The aim of the workshop is to answer cutting edge questions about dark matter with the most advanced techniques that data science provides us with. The researchers coming to this workshop will bring together expertise in particle physics, astrophysics, cosmology, statistics and information science (machine learning) and plenty of time will be devoted to discussions and brainstorming.


Workshop Format
 
The workshop will have plenary sessions covering the main topics in the field of machine learning and dark matter. In addition, there will be a limited number of parallel talks. We encourage scientists interested to present their work on applications of machine learning to dark matter problems to submit their abstracts. 

Afternoons will be devoted to hands-on sessions, during which the participants will work on the ongoing darkmachines challenges (listed here http://darkmachines.org/#projects) or propose and start new ones.

The last day of the meeting will be dedicated to the discussion related to the DarkMachines organisational issues (progress of the white paper, summary of computational challenges etc).
 

Invited Speakers include:

Jelle Aalbers (Stockholm University)
Atılım Güneş Baydin (University of Oxford)
Johann Cohen-Tanugi (University of Montpellier)
Caterina Doglioni (Lund University)
Torsten Ensslin (Max Planck Institute for Astrophysics)
Guðlaugur Jóhannesson (University of Iceland)
Jaehoon Lee (Google Brain)
Nikita Moriakov (Radboud University Medical Center)
Daniel Nieto (Universidad Complutense de Madrid)
Rajat Thomas (University of Amsterdam)
Danilo J. Rezende (DeepMind, London)
Pasquale Serpico (LAPTH, Annecy)
David Shih (Rutgers University)
Roberto Trotta (Imperial College, London)
Simona Vegetti (Max Planck Institute for Astrophysics )


Coordinators of Dark Machines Challenges include:

Amir Farbin (UT Arlington)
Luc Hendriks (Radboud University)
Fabio Iocco (ICTP-SAIFR)
Erzsébet Merényi (Rice University)
Michela Negro (INFN Torino)
Sydney Otten (Radboud University)
Maurizio Pierini (CERN)
Christoph Weniger (GRAPPA,  Amsterdam)
Martin White (University of Adelaide)


Scientific Organizing Committee:

Gianfranco Bertone (GRAPPA, Amsterdam)
Francesca Calore (LAPTH, Annecy)
Sascha Caron (NIKEF, Nijmegen)
Tommaso Dorigo (University of Padova)  
Gilles Louppe (University of Liège)
Roberto Ruiz de Austri (IFIC, Paterna)


NOTE: The deadline for requesting participation was 10 March 2019. For any queries please e-mail smr3280@ictp.it.

There is no registration fee for this activity.
Go to day
  • Monday, 8 April 2019
    • 08:30 - 18:30 Monday, 8 April 2019 - INDIRECT DETECTION
      Location: Adriatico Guest House - Kastler Lecture Hall
      • 08:30 Registration & Administrative Formalities 20' ( Adriatico Guest House - Workshop Secretariat )
      • 08:50 Welcome address 10'
        Material: Slides
      • 09:00 Overview of indirect DM search challenges 45'
        Speaker: Pasquale Serpico (LAPTH, Annecy)
        Material: Slides
      • 09:45 Probabilistic programming 45'
        Speaker: Rajat Thomas (University of Amsterdam)
        Material: Slides
      • 10:30 Coffee Break 30' ( Adriatico Guest House - Terrace )
      • 11:00 Fermi-LAT sky 45'
        Speaker: Guðlaugur Jóhannesson (University of Iceland)
        Material: Slides
      • 11:45 Computer vision 45'
        Speaker: Nikita Moriakov (Radboud University Medical Center)
      • 12:30 Lunch Break 1h45' ( Adriatico Guest House - Cafeteria )
      • 14:15 ML and cosmic ray shower reconstruction with CTA 30'
        Speaker: Daniel Nieto (Universidad Complutense de Madrid)
        Material: Slides
      • 14:45 New data challenge on deep probabilistic programming 30'
        Speaker: Christoph Weniger (GRAPPA, Amsterdam)
      • 15:15 Status of challenges + Brainstorming 45'
        Material: Slides
      • 16:00 Coffee Break 20' ( Adriatico Guest House - Terrace )
      • 16:20 Work in groups 1h0'
      • 17:30 IFPU lectures - Elements of ML (= statistical learning) theory 1h0'
        GitHub page
        Speaker: Gilles Louppe (University of Liège)
    • 19:00 - 19:01 Dinner Reception
      Location: Adriatico Guest House - Cafeteria
  • Tuesday, 9 April 2019
    • 09:00 - 18:30 Tuesday, 9 April 2019 - PARTICLE PHYSICS
      Location: Adriatico Guest House - Kastler Lecture Hall
      • 09:00 Searches for DM at the Large Hadron Collider 45'
        Speaker: Caterina Doglioni (Lund University)
        Material: Slides
      • 09:45 Probabilistic Programming and Inference in Particle Physics 45'
        Speaker: Atılım Güneş Baydin (University of Oxford))
        Material: Slides
      • 10:30 Coffee Break 30' ( Adriatico Guest House - Terrace )
      • 11:00 Boosting New Physics Searches with Machine Learning 45'
        Speaker: David Shih (Rutgers University)
        Material: Slides
      • 11:45 Magnifying (unknown) rare clusters to increase the chance of detection, using unsupervised learning 30'
        Speaker: Erzsébet Merényi (Rice University)
        Material: Slides
      • 12:15 Variational Autoencoders 20'
        Speaker: Luc Hendriks (Radboud University)
        Material: Slides
      • 12:35 Lunch Break 1h45' ( Adriatico Guest House - Cafeteria )
      • 14:20 Event Generation with Deep Generative Models 20'
        Speaker: Sydney Otten (Radboud University)
      • 14:40 Recent results from the GAMBIT collaboration 20'
        Speaker: Martin White (University of Adelaide)
        Material: Slides
      • 15:00 Collider searches for Dark Matter at LHC using machine learning 20'
        Speaker: Charanjit Kaur & Michael Soughton (University of Sussex)
        Material: Slides
      • 15:20 Brainstorming + Status of challenges 40'
        Material: Slides
      • 16:00 Coffee Break 20' ( Adriatico Guest House - Terrace )
      • 16:20 Work in groups 1h0'
      • 17:30 IFPU lectures - Neural networks (from logistic regression to deep networks) 1h0' ( IFPU, via Beirut 2 )
        Speaker: Gilles Louppe (University of Liège)
  • Wednesday, 10 April 2019
    • 09:00 - 19:10 Wednesday, 10 April 2019 - GRAVITATIONAL LENSING
      Location: Adriatico Guest House - Kastler Lecture Hall
      • 09:00 Overview of the LSST DM program 45'
        Speaker: Johann Cohen-Tanugi (University of Montpellier)
        Material: Slides
      • 09:45 Deep Generative Models: Foundations, applications and open problems (Video Call) 45'
        Speaker: Danilo J. Rezende (DeepMind, London)
      • 10:30 Coffee Break 30' ( Adriatico Guest House - Terrace )
      • 11:00 DM and gravitational lensing 45'
        Speaker: Simona Vegetti (Max Planck Institute for Astrophysics)
      • 11:45 Strong Gravitational Lensing and ML: Methodology 15'
        Speaker: Adam Coogan (GRAPPA, Amsterdam)
        Material: Slides
      • 12:00 Strong Gravitational Lensing Systems with Machine Learning: first results on Dark Matter substructures 15'
        Speaker: Marco Chianese (GRAPPA, Amsterdam)
        Material: Slides
      • 12:15 Lunch Break 1h45' ( Adriatico Guest House - Cafeteria )
      • 14:00 Determining the dark matter clumps parameters from strong gravitational microlensing events 15'
        Speaker: Olena Fedorova (University of Kyiv)
      • 14:15 Search for dark substructure in Gaia stellar field using CNN 15'
        Speaker: Mihael Petac (SISSA)
        Material: Slides
      • 14:30 Using machine learning to unlock Gaia's full potential to determine the dark matter halo 15' ( Adriatico Guest House - Kastler Lecture Hall )
        Speaker: Bryan Ostdiek (University of Oregon)
      • 15:00 EHT Press Conference 1h0'
        Link
      • 16:00 Coffee Break 30' ( Adriatico Guest House - Terrace )
      • 16:30 Gaussian processes and deep learning (Video Call) 45'
        Speaker: Jaehoon Lee (Google Brain)
        Material: Slides
      • 17:15 Brainstorming + Status of challenges 45'
        Material: Slides
      • 18:00 Work in groups (TBC) 1h0' ( Adriatico Guest House - Kastler Lecture Hall )
    • 20:00 - 23:00 Workshop Dinner
      • 20:00 Ristorante Menarosti, Via del Toro 12, Trieste 3h0' ( Trieste ( City Center ) )
        Reserved for Workshop Participants only. Confirmation is mandatory by e-mailing smr3280@ictp.it
        Material: Directions & Tips
  • Thursday, 11 April 2019
    • 09:00 - 17:50 Thursday, 11 April 2019 - DIRECT DETECTION
      Location: Adriatico Guest House - Kastler Lecture Hall
      • 09:00 Overview of direct detection challenges 45'
        Speaker: Jelle Aalbers (Stockholm University)
        Material: Slides
      • 09:45 Information field theory 45'
        Speaker: Torsten Ensslin (Max Planck Institute for Astrophysics)
        Material: Slides
      • 10:30 Coffee Break 30' ( Adriatico Guest House - Terrace )
      • 11:00 Bayesian Tools for Dark Matter Searches 45'
        Speaker: Roberto Trotta (Imperial College, London)
        Material: Slides
      • 11:45 Likelihood-free Markov chain Monte Carlo with Approximate Likelihood Ratios 20'
        Speaker: Joeri Hermans (University of Liège)
      • 12:05 Noisy-input classification of Fermi-LAT unidentified point-like sources 15'
        Speaker: Bryan Zaldivar Montero (IFT, Madrid)
        Material: Slides
      • 12:20 DM & CTA 15'
        Speaker: Salvatore Mangano (Universidad Politécnica de Madrid)
        Material: Slides
      • 12:35 Lunch Break 1h45' ( Adriatico Guest House - Cafeteria )
      • 14:20 An application of unsupervised learning method in HEP data analysis 15'
        Speaker: Ying-Ying Li (Hong Kong U. of Science and Tech.)
        Material: Slides
      • 14:35 Tagging jets with LDA topic models 15'
        Speaker: Barry Dillon (Jozef Stefan Institute, Ljubljana)
        Material: Slides
      • 14:50 Search for patterns in physics: from particles to dark matter 15'
        Speaker: Josue Molina Rodriguez (Universidad Zamorano)
        Material: Slides
      • 15:05 Elementary particle based models for Dark Matter on galaxy scales 15'
        Speaker: Carlos Argüelles (UNLP & CONICET)
        Material: Slides
      • 15:20 Bayesian reconstruction of the Milky Way dark matter distribution 15' ( Adriatico Guest House - Kastler Lecture Hall )
        Speaker: Ekaterina Karukes (CAMK/AstroCeNT, Warsaw)
        Material: Slides
      • 15:35 Status of challenges + Brainstorming 45' ( Adriatico Guest House - Kastler Lecture Hall )
      • 16:20 Coffee Break 30' ( Adriatico Guest House - Terrace )
      • 16:50 Work in groups 1h0' ( Adriatico Guest House - Kastler Lecture Hall )
  • Friday, 12 April 2019
    • 09:00 - 12:00 Friday, 12 April 2019
      Location: Adriatico Guest House - Kastler Lecture Hall
      • 09:00 Summary of darkmachines.org challenges 1h30'
      • 10:30 Coffee Break 30' ( Adriatico Guest House - Terrace )
      • 11:00 Next steps 1h0'