Inverse Reinforcement Learning with Natural Language Goals
This work addresses the problem of reducing friction in task specification for robots or agents by using natural language, representing an incremental advance in generalization for language-conditioned policies.
The paper tackles the challenge of enabling autonomous machines to understand and act on natural language goals by proposing an adversarial inverse reinforcement learning algorithm with a variational goal generator, achieving a large margin improvement over baselines on the Room-2-Room dataset.
Humans generally use natural language to communicate task requirements to each other. Ideally, natural language should also be usable for communicating goals to autonomous machines (e.g., robots) to minimize friction in task specification. However, understanding and mapping natural language goals to sequences of states and actions is challenging. Specifically, existing work along these lines has encountered difficulty in generalizing learned policies to new natural language goals and environments. In this paper, we propose a novel adversarial inverse reinforcement learning algorithm to learn a language-conditioned policy and reward function. To improve generalization of the learned policy and reward function, we use a variational goal generator to relabel trajectories and sample diverse goals during training. Our algorithm outperforms multiple baselines by a large margin on a vision-based natural language instruction following dataset (Room-2-Room), demonstrating a promising advance in enabling the use of natural language instructions in specifying agent goals.