Scientific Calendar Event



Description
 
The video recordings of the lectures and seminars held during the School can be viewed at this link:
https://www.youtube.com/playlist?list=PLRwcSE2bmyBwzkJzLAT__C7h_FaweUqDL

Life can be thought of as a type of stochastic information processing realized out of thermodynamic equilibrium. Living systems indeed consume energy to sustain order and information flow in fluctuating environments. Recent advances in non-equilibrium statistical physics and information theory offer a new framework to unravel the underlying principles of the dynamical processes in living matter. As an example of the developing interface between biology, information theory and statistical physics, a recently discovered new principle, the thermodynamic uncertainty relation, quantifies the energy-precision trade-off of biological processes and its physical bounds. Another important example of the richness of this interface is offered by machine learning, which couples ideas from neuroscience, computer science and statistical physics.

The new language developed in modern statistical mechanics is then expected to play a key role in the understanding of how order, and eventually life, emerges from disorder. It has been widely demonstrated that statistical mechanics is an essential tool for the development of the above subjects, and great interest in these subjects is currently shared by statistical physics and biophysics communities in Korea.
The 2022 ICTP-KIAS School on Statistical Physics for Life Sciences aims to introduce the next generation of students and researchers to these important themes: the quantitative understanding of life sciences, machine learning, and data science from the perspective of modern statistical physics. In addition to the pedagogical lectures, the school includes a small workshop for sharing recent research progresses among scientists at the frontier of this field.

The school aims at providing unique opportunities for (junior) scientists in the Asia Pacific region to learn modern statistical physics and its applications to quantitative life science, machine learning, and data science. The 2022 ICTP-KIAS School will contribute to the cross-fertilization between physics and biology as well as to long-term scientific collaborations between researchers in Asia and ICTP scientists.

Topics

Information theory, machine learning and statistical physics
  • Statistical mechanics of machine learning
  • Information theory and criticality
  • Statistical mechanics and reasoning

Out-of-equilibrium statistical physics and life
  • Biochemical rhythms and thermodynamics
  • Non-equilibrium thermodynamics
  • Information thermodynamics
  • Stochastic thermodynamics
Lecturers for the School:
 
Vipul Periwal (National Institutes of Health)
Jae Sung Lee (Korea Institute for Advance Study)
Pankaj Mehta (Boston University)
Takahiro Sagawa (University of Tokyo)

Speakers for the workshop:
 
Changbong Hyeon (Korea Institute for Advance Study)
Hyunggyu Park (Korea Institute for Advance Study)
Yongjoo Baek (Seoul National University)
Tetsuya Kobayashi (University of Tokyo)
Jae-Hyung Jeon (Pohang University of Science and Technology)
Yong Woon Kim (Korea Advanced Institute of Science and Technology)
Kunihiko Kaneko (Niels Bohr Institute)
Antonio Celani (The Abdus Salam International Centre for Theoretical Physics)
Vipul Periwal (National Institutes of Health)
Deok-Sun Lee (Korea Institute for Advance Study)
Junghyo Jo (Seoul National University)


 






 

 

 
Go to day
  • Monday, 31 October 2022
    • 09:30 - 17:15 Monday, 31 October
      • 09:30 Registration 3h45'
      • 12:15 Lunch break 2h15'
      • 14:30 Comparing the inverse problem in statistical physics to model selection in data science 1h15'
        Speaker: Vipul Periwal, NIH
      • 15:45 Break 15'
      • 16:00 Work, Heat, and Stochastic Trajectory in Stochastic Systems 1h15'
        Speaker: Jae Sung Lee, KIAS
  • Tuesday, 1 November 2022
    • 09:30 - 17:15 Tuesday, 1 November
      • 09:30 Kinetic Ising models for modeling discrete data 1h15'
        Speaker: Vipul Periwal, NIH
      • 10:45 Break 15'
      • 11:00 Entropy Production and Fluctuation Theorems 1h15'
        Speaker: Jae Sung Lee, KIAS
      • 12:15 Lunch break 2h15'
      • 14:30 Discussion 1h30'
      • 16:00 Thermodynamic Uncertainty Relations 1h15'
        Speaker: Jae Sung Lee, KIAS
  • Wednesday, 2 November 2022
    • 09:30 - 17:15 Wednesday, 2 November
      • 09:30 Fitting versus Predicting 1h15'
        Speaker: Pankaj Mehta, Boston U.
      • 10:45 Break 15'
      • 11:00 Temperature in data science 1h15'
        Speaker: Vipul Periwal, NIH
      • 12:15 Lunch break 2h15'
      • 14:30 Discussion 1h30'
      • 16:00 Thermodynamic Speed Limit 1h15'
        Speaker: Jae Sung Lee, KIAS
  • Thursday, 3 November 2022
    • 09:30 - 17:15 Thursday, 3 November
      • 09:30 Some new approaches to old problems 1h15'
        Speaker: Vipul Periwal, NIH
      • 10:45 Break 15'
      • 11:00 Towards Deep Learning: Linear and Logistic Regression 1h15'
        Speaker: Pankaj Mehta, Boston U.
      • 12:15 Lunch break 2h15'
      • 14:30 An introduction to information theory 1h15'
        Speaker: Takahiro Sagawa, Tokyo Institute of Technology
      • 15:45 Break 15'
      • 16:00 Thermodynamics of information I 1h15'
        Speaker: Takahiro Sagawa, Tokyo Institute of Technology
  • Friday, 4 November 2022
    • 09:30 - 17:15 Friday, 4 November
      • 09:30 Introduction to modern Deep Learning 1h15'
        Speaker: Pankaj Mehta, Boston U.
      • 10:45 Break 15'
      • 11:00 Why does modern Deep Learning work so well? Rethinking generalization 1h15'
        Speaker: Pankaj Mehta, Boston U.
      • 12:15 Lunch break 2h15'
      • 14:30 Thermodynamics of information II 1h15'
        Speaker: Takahiro Sagawa, Tokyo Institute of Technology
      • 16:00 Autonomous Maxwell's demons 1h15'
        Speaker: Takahiro Sagawa, Tokyo Institute of Technology
  • Monday, 7 November 2022
    • 09:30 - 18:30 Monday, 7 November
      • 09:30 Molecular chaperones as nonequilibrium annealing machines 30'
        Speaker: Changbong Hyeon (KIAS)
      • 10:10 On the origin of the power-efficiency bound of a heat engine 30'
        Speaker: Hyunggyu Park (KIAS)
      • 10:50 A generic mechanism for symmetry-breaking motility in a dilute active fluid 30'
        Speaker: Yongjoo Baek (Seoul National University)
      • 11:30 Molecular mechanisms for filtering spatio-temporal noise in cells (ONLINE) 30'
        Speaker: Jae Kyoung Kim (KAIST/IBS)
      • 12:00 Lunch break 2h0'
      • 14:00 Chemical Thermodynamics of Growing Systems 30'
        Speaker: Tetsuya Kobayashi (University of Tokyo)
      • 14:40 Active diffusion of self-propelled particles interacting with a polymer network 30'
        Speaker: Jae-Hyung Jeon (Pohang University of Science and Technology)
      • 15:20 First-passage dynamics of many Brownian particles 30'
        Speaker: Yong Woon Kim (Korea Advanced Institute for Science and Technology)
      • 16:00 Dynamical systems theory of robust cell differentiation and reprogramming (ONLINE) 30'
        Speaker: Kunihiko Kaneko (University of Tokyo)
      • 16:50 Poster session 1h40'
  • Tuesday, 8 November 2022
    • 09:30 - 14:00 Tuesday, 8 November
      • 09:30 The uses of memory in olfactory search 30'
        Speaker: Antonio Celani (ICTP)
      • 10:10 Scaling persistent homology to large biological datasets 30'
        Speaker: Vipul Periwal (NIH)
      • 10:50 Evolution of metabolic reactions’ heterogeneous popularity 30'
        Speaker: Deok-Sun Lee (KIAS)
      • 11:30 Mirror descent of Hopfield model 30'
        Speaker: Junghyo Jo (Seoul National University)
      • 12:00 Lunch break 2h0'