DeepAI AI Chat
Log In Sign Up

Graph networks as learnable physics engines for inference and control

by   Alvaro Sanchez-Gonzalez, et al.

Understanding and interacting with everyday physical scenes requires rich knowledge about the structure of the world, represented either implicitly in a value or policy function, or explicitly in a transition model. Here we introduce a new class of learnable models--based on graph networks--which implement an inductive bias for object- and relation-centric representations of complex, dynamical systems. Our results show that as a forward model, our approach supports accurate predictions from real and simulated data, and surprisingly strong and efficient generalization, across eight distinct physical systems which we varied parametrically and structurally. We also found that our inference model can perform system identification. Our models are also differentiable, and support online planning via gradient-based trajectory optimization, as well as offline policy optimization. Our framework offers new opportunities for harnessing and exploiting rich knowledge about the world, and takes a key step toward building machines with more human-like representations of the world.


page 1

page 3

page 4

page 5

page 7

page 20


Propagation Networks for Model-Based Control Under Partial Observation

There has been an increasing interest in learning dynamics simulators fo...

Neural Relational Inference for Interacting Systems

Interacting systems are prevalent in nature, from dynamical systems in p...

Reasoning About Physical Interactions with Object-Oriented Prediction and Planning

Object-based factorizations provide a useful level of abstraction for in...

Visual Interaction Networks

From just a glance, humans can make rich predictions about the future st...

Trajectory Optimization Using Neural Network Gradients of Learned Dynamics

Trajectory optimization methods have achieved an exceptional level of pe...

Learning Generalized Quasi-Geostrophic Models Using Deep Neural Numerical Models

We introduce a new strategy designed to help physicists discover hidden ...