Timezone: »

 
Poster
Structured agents for physical construction
Victor Bapst · Alvaro Sanchez-Gonzalez · Carl Doersch · Kimberly Stachenfeld · Pushmeet Kohli · Peter Battaglia · Jessica Hamrick

Wed Jun 12 06:30 PM -- 09:00 PM (PDT) @ Pacific Ballroom #36
in Posters Wed »

Physical construction---the ability to compose objects, subject to physical dynamics, to serve some function---is fundamental to human intelligence. We introduce a suite of challenging physical construction tasks inspired by how children play with blocks, such as matching a target configuration, stacking blocks to connect objects together, and creating shelter-like structures over target objects. We examine how a range of deep reinforcement learning agents fare on these challenges, and introduce several new approaches which provide superior performance. Our results show that agents which use structured representations (e.g., objects and scene graphs) and structured policies (e.g., object-centric actions) outperform those which use less structured representations, and generalize better beyond their training when asked to reason about larger scenes. Model-based agents which use Monte-Carlo Tree Search also outperform strictly model-free agents in our most challenging construction problems. We conclude that approaches which combine structured representations and reasoning with powerful learning are a key path toward agents that possess rich intuitive physics, scene understanding, and planning.

Keywords: Deep Reinforcement Learning · Networks and Relational Learning · Other Applications · Planning and Control

Author Information

Victor Bapst (Google DeepMind)
Alvaro Sanchez-Gonzalez (DeepMind)
Carl Doersch (DeepMind)
Kimberly Stachenfeld (Google)
Pushmeet Kohli (DeepMind)
Peter Battaglia (DeepMind)
Jessica Hamrick (DeepMind)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors