Workshop: Interpretable Inductive Biases and Physically Structured Learning

Michael Lutter, Alexander Terenin, Shirley Ho, Lei Wang

2020-12-12T06:30:00-08:00 - 2020-12-12T14:30:00-08:00
Abstract: Over the last decade, deep networks have propelled machine learning to accomplish tasks previously considered far out of reach, human-level performance in image classification and game-playing. However, research has also shown that the deep networks are often brittle to distributional shifts in data: it has been shown that human-imperceptible changes can lead to absurd predictions. In many application areas, including physics, robotics, social sciences and life sciences, this motivates the need for robustness and interpretability, so that deep networks can be trusted in practical applications. Interpretable and robust models can be constructed by incorporating prior knowledge within the model or learning process as an inductive bias, thereby regularizing the model, avoiding overfitting, and making the model easier to understand for scientists who are non-machine-learning experts. Already in the last few years researchers from different fields have proposed various combinations of domain knowledge and machine learning and successfully applied these techniques to various applications.

Video

Chat

Chat is not available.

Schedule

2020-12-12T06:30:00-08:00 - 2020-12-12T06:35:00-08:00
Introduction
2020-12-12T06:35:00-08:00 - 2020-12-12T06:50:00-08:00
Thomas Pierrot - Learning Compositional Neural Programs for Continuous Control
Thomas PIERROT
2020-12-12T06:50:00-08:00 - 2020-12-12T07:10:00-08:00
Jessica Hamrick - Structured Computation and Representation in Deep Reinforcement Learning
Jessica Hamrick
2020-12-12T07:10:00-08:00 - 2020-12-12T07:25:00-08:00
Manu Kalia - Deep learning of normal form autoencoders for universal, parameter-dependent dynamics
Manu Kalia
2020-12-12T07:25:00-08:00 - 2020-12-12T07:50:00-08:00
Rose Yu - Physics-Guided AI for Learning Spatiotemporal Dynamics
Rose Yu
2020-12-12T07:50:00-08:00 - 2020-12-12T08:05:00-08:00
Ferran Alet - Tailoring: encoding inductive biases by optimizing unsupervised objectives at prediction time
Ferran Alet
2020-12-12T08:05:00-08:00 - 2020-12-12T09:00:00-08:00
Poster Session 1
2020-12-12T09:00:00-08:00 - 2020-12-12T09:25:00-08:00
Frank Noé - PauliNet: Deep Neural Network Solution of the Electronic Schrödinger Equation
Frank Noe
2020-12-12T09:25:00-08:00 - 2020-12-12T09:40:00-08:00
Kimberly Stachenfeld - Graph Networks with Spectral Message Passing
Kim Stachenfeld
2020-12-12T09:40:00-08:00 - 2020-12-12T10:10:00-08:00
Franziska Meier - Inductive Biases for Models and Learning-to-Learn
Franziska Meier
2020-12-12T10:10:00-08:00 - 2020-12-12T10:25:00-08:00
Rui Wang - Shapley Explanation Networks
Rui Wang
2020-12-12T10:25:00-08:00 - 2020-12-12T10:55:00-08:00
Jeanette Bohg - One the Role of Hierarchies for Learning Manipulation Skills
Christin Jeannette Bohg
2020-12-12T11:00:00-08:00 - 2020-12-12T12:00:00-08:00
Panel Discussion
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
4 - Physics-informed Generative Adversarial Networks for Sequence Generation with Limited Data
Chacha Chen
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
25 - Complex Skill Acquisition through Simple Skill Imitation Learning
Pranay Pasula
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
12 - IV-Posterior: Inverse Value Estimation forInterpretable Policy Certificates
Tatiana López-Guevara
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
9 - Thermodynamic Consistent Neural Networks for Learning Material Interfacial Mechanics
Jiaxin Zhang
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
20 -SOrT-ing VQA Models : Contrastive Gradient Learning for Improved Consistency
Sameer Dharur
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
5 - On the Structure of Cyclic Linear Disentangled Representations
Matthew Painter
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
Poster Session 2
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
26 - Is the Surrogate Model Interpretable?
Sangwon Kim
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
6 - Interpretable Models for Granger Causality Using Self-explaining Neural Networks
Ričards Marcinkevičs
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
21 - Solving Physics Puzzles by Reasoning about Paths
Augustin Harter
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
19 - Choice of Representation Matters for Adversarial Robustness
Amartya Sanyal
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
17 - Uncovering How Neural Network Representations Vary with Width and Depth
Thao Nguyen
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
24 - Deep Context-Aware Novelty Detection
Ellen Rushe
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
14 - Learning Dynamical Systems Requires Rethinking Generalization
Rui Wang
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
10 - A Trainable Optimal Transport Embedding for Feature Aggregation
Grégoire Mialon
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
1 - Real-time Classification from Short Event-Camera Streams using Input-filtering Neural ODEs
Giorgio Giannone
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
11 - A novel approach for semiconductor etching process with inductive biases
Sanghoon Myung
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
23 - Constraining neural networks output by an interpolating loss function with region priors
Hannes Bergkvist
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
8 - Individuality in the hive - Learning to embed lifetime social behavior of honey bees
Benjamin Wild
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
3 - Improving the trustworthiness of image classification models by utilizing bounding-box annotations
Dharma R KC
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
15 - Lie Algebra Convolutional Networks with Automatic Symmetry Extraction
Nima Dehmamy
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
7 - A Symmetric and Object-Centric World Model for Stochastic Environments
Patrick Emami
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
2 - Relevance of Rotationally Equivariant Convolutions for Predicting Molecular Properties
Benjamin K Miller
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
13 - Gradient-based Optimization for Multi-resource Spatial Coverage
Nitin Kamra
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
12 - Physics-aware, data-driven discovery of slow and stable coarse-grained dynamics for high-dimensional multiscale systems
Sebastian Kaltenbach
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
22 - Modelling Advertising Awareness, an Interpretable and Differentiable Approach
Luz Blaz
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
18 - Simulating Surface Wave Dynamics with Convolutional Networks
2020-12-12T12:00:00-08:00 - 2020-12-12T13:00:00-08:00
16 - An Image is Worth 16 × 16 Tokens: Visual Priors for Efficient Image Synthesis with Transformers
Robin Rombach
2020-12-12T13:00:00-08:00 - 2020-12-12T13:15:00-08:00
Liwei Chen - Deep Learning Surrogates for Computational Fluid Dynamics
Nils Thuerey
2020-12-12T13:15:00-08:00 - 2020-12-12T14:15:00-08:00
Maziar Raissi - Hidden Physics Models
Maziar Raissi
2020-12-12T14:15:00-08:00 - 2020-12-12T14:30:00-08:00
Closing Remarks