Model-based language-instructed reinforcement learning

Anonymous

Abstract: We explore how we can build accurate world models which are partially specified by language and how we can plan with them in the face of novelty and uncertainty. We propose the first Model-Based Reinforcement Learning approach to tackle the environment Read To Fight Monsters (Zhong et al., 2019), a grounded policy learning problem. In RTFM an agent has to reason over a set of rules and a goal, both described in a language manual, and the observations, while taking into account the uncertainty arising from the stochasticity of the environment, in order to generalize successfully its policy to test episodes. We provide a sample-efficient proof-of-concept of the model-based approach for the basic dynamic task of RTFM.Furthermore, we show that the main open challenge of RTFM is learning the language-dependent reward function and suggest that future research should focus primarily on that task.

Paper Type: long

Research Area: Multimodality and Language Grounding to Vision, Robotics and Beyond