Workshop II: Interpretable Learning in Physical Sciences - IPAM

Workshop II: Interpretable Learning in Physical Sciences

Part of the Long Program Machine Learning for Physics and the Physics of Learning
October 14 - 18, 2019

Schedule

All times in this Schedule are Pacific Time (PT)

Monday, October 14, 2019

Morning Session

08:00-08:55
Check-In/Light Breakfast (Hosted by IPAM)
08:55-09:00
Welcome & Opening Remarks: Dean Miguel García-Garibay (Dean of Physical Sciences, UCLA) and Dima Shlyakhtenko (Director, IPAM)
09:00-09:50
Cecilia Clementi (Freie Universität Berlin)
Learning molecular model from simulation and experimental data
10:00-10:15
Break
10:15-11:05
Yannis Kevrekides (Princeton University)
No equations, no variables, no space and no time: data driven models and gauge-invariant data mining
11:15-11:30
Break
11:30-12:20
Michael Mahoney (University of California, Berkeley (UC Berkeley))
Why Deep Learning Works: Heavy-Tailed Random Matrix Theory as an Example of Physics Informed Machine Learning

Afternoon Session

12:30-14:30
Lunch (on your own)
14:30-15:20
Shirley Ho (Flatiron Institute, a Division of the Simons Foundation)
Simulating the Universe with Deep Learning
15:30-15:45
Break
15:45-16:35
Julia Ling (Citrine Informatics)
Model Interpretability for Building Confidence and Sparking Insight in Scientific Applications
16:45-17:15
Lightning Poster Presentations
17:15-18:30
Poster Session & Reception (Hosted by IPAM)

Tuesday, October 15, 2019

Morning Session

08:00-09:00
Check-In/Breakfast (Hosted by IPAM)
09:00-09:50
Weinan E (Princeton University)
Machine learning based multi-scale modeling
10:00-10:15
Break
10:15-11:05
Andrew White (University of Rochester)
Maximum Entropy Methods for Combining Physics-Based Simulation with Empirical Data
11:15-11:30
Break
11:30-12:20
Heather Kulik (Massachusetts Institute of Technology)
Molecular design blueprints: materials and catalysts from new simulation and machine learning tools

Afternoon Session

12:30-14:30
Lunch (on your own)
14:30-15:20
Atilim Gunes Baydin (University of Oxford)
Universal Probabilistic Programming in Simulators
15:30-16:00
Break
16:00-16:50
Mark Waller (Wallerlab)
Planning Chemical Syntheses with Deep Neural Networks and Discipline Scale Data

Wednesday, October 16, 2019

Morning Session

08:00-09:00
Check-In/Breakfast (Hosted by IPAM)
09:00-09:50
J. Nathan Kutz (University of Washington)
Coordinates, governing equations and limits of model discovery
10:00-10:15
Break
10:15-11:05
Peter Frazier (Cornell University)
Accelerating Scientific Discovery through Interpretable Machine Learning and Intelligent Experimentation
11:15-11:30
Break
11:30-12:20
Peter Battaglia (DeepMind Technologies)
Learning structured models of physics

Afternoon Session

12:30-14:00
Lunch (on your own)
14:00-14:50
Judea Pearl (University of California, Los Angeles (UCLA))
Interpretability and explainability from a causal lens
15:00-15:15
Break
15:15-16:05
Adji Bousso Dieng (Columbia University)
Structured Deep Generative Models
16:15-16:30
Break
16:30-17:20
Gianni De Fabritiis (Universitat Pompeu Fabra)
Can we machine learn drug discovery?

Thursday, October 17, 2019

Morning Session

08:00-09:00
Check-In/Breakfast (Hosted by IPAM)
09:00-09:50
Sergei Kalinin (University of Tennessee)
Deep Learning Dive into the Scanning Transmission Electron Microscopy: Materials Design, Learning Physics, and Atomic Manipulation
10:00-10:15
Break
10:15-11:05
Anatole von Lilienfeld (University of Toronto)
Quantum Machine Learning
11:15-11:30
Break
11:30-12:20
Kieron Burke (University of California, Irvine (UCI))
Density functionals from machine learning

Afternoon Session

12:30-14:30
Lunch (on your own)
14:30-15:20
Kyle Cranmer (University of Wisconsin-Madison)
Simulation-based inference, interpretability, and experimental design
15:30-16:00
Break
16:00-16:50
Zachary Lipton (Carnegie Mellon University)
Interpretability: of what, for whom, why, and how?

Friday, October 18, 2019

Morning Session

08:00-09:00
Check-In/Breakfast (Hosted by IPAM)
09:00-09:50
Stephan Hoyer (Google Inc.)
Improving PDE solvers and PDE-constrained optimization with deep learning and differentiable programming
10:00-10:15
Break
10:15-11:05
Pratyush Tiwary (University of Maryland)
Learning to learn, learning to forget
11:15-11:30
Break
11:30-12:20
Andrew Ferguson (University of Chicago)
Machine learning of protein folding funnels from experimental data and construction of latent space molecular simulators

Afternoon Session

12:30-14:30
Lunch (on your own)
14:30-15:20
Francesco Paesani (University of California, San Diego (UCSD))
Data-driven models for predictive molecular simulations
15:30-16:00
Break
16:00-16:50
Tristan Bereau (Max Planck Institute for Polymer Research)
Physics in and out of machine learning for molecular simulations: priors and predictive constraints
Printable Version