Learning to Understand Goal Specifications by Modelling Reward
This addresses the problem of scaling language-based instructions in complex environments for AI researchers, though it is incremental as it builds on existing instruction-conditional RL methods.
The paper tackles the challenge of designing language-conditional reward functions for reinforcement learning agents by introducing a framework where agents learn from reward models trained on expert examples, enabling them to follow instructions in unseen configurations and adapt to environmental changes without new data.
Recent work has shown that deep reinforcement-learning agents can learn to follow language-like instructions from infrequent environment rewards. However, this places on environment designers the onus of designing language-conditional reward functions which may not be easily or tractably implemented as the complexity of the environment and the language scales. To overcome this limitation, we present a framework within which instruction-conditional RL agents are trained using rewards obtained not from the environment, but from reward models which are jointly trained from expert examples. As reward models improve, they learn to accurately reward agents for completing tasks for environment configurations---and for instructions---not present amongst the expert data. This framework effectively separates the representation of what instructions require from how they can be executed. In a simple grid world, it enables an agent to learn a range of commands requiring interaction with blocks and understanding of spatial relations and underspecified abstract arrangements. We further show the method allows our agent to adapt to changes in the environment without requiring new expert examples.