Grounding Language to Entities and Dynamics for Generalization in Reinforcement Learning

Austin W. Hanjie · Victor Zhong · Karthik Narasimhan

[ Abstract ] [ Livestream: Visit Reinforcement Learning 11 ] [ Paper ]
Wed 21 Jul 5:40 a.m. — 5:45 a.m. PDT

We investigate the use of natural language to drive the generalization of control policies and introduce the new multi-task environment Messenger with free-form text manuals describing the environment dynamics. Unlike previous work, Messenger does not assume prior knowledge connecting text and state observations — the control policy must simultaneously ground the game manual to entity symbols and dynamics in the environment. We develop a new model, EMMA (Entity Mapper with Multi-modal Attention) which uses an entity-conditioned attention module that allows for selective focus over relevant descriptions in the manual for each entity in the environment. EMMA is end-to-end differentiable and learns a latent grounding of entities and dynamics from text to observations using only environment rewards. EMMA achieves successful zero-shot generalization to unseen games with new dynamics, obtaining a 40% higher win rate compared to multiple baselines. However, win rate on the hardest stage of Messenger remains low (10%), demonstrating the need for additional work in this direction.

Chat is not available.