Rethinking Inverse Reinforcement Learning: from Data Alignment to Task Alignment
This work addresses task-reward misalignment in imitation learning, which is a critical issue for improving the robustness and transferability of learned policies in robotics and AI applications, though it appears incremental as it builds on existing IRL methods.
The paper tackles the problem of inverse reinforcement learning (IRL) in imitation learning, where inferred reward functions often misalign with task objectives, by proposing a novel framework that prioritizes task alignment over data alignment, resulting in outperformance of conventional baselines in complex and transfer learning scenarios.
Many imitation learning (IL) algorithms use inverse reinforcement learning (IRL) to infer a reward function that aligns with the demonstration. However, the inferred reward functions often fail to capture the underlying task objectives. In this paper, we propose a novel framework for IRL-based IL that prioritizes task alignment over conventional data alignment. Our framework is a semi-supervised approach that leverages expert demonstrations as weak supervision to derive a set of candidate reward functions that align with the task rather than only with the data. It then adopts an adversarial mechanism to train a policy with this set of reward functions to gain a collective validation of the policy's ability to accomplish the task. We provide theoretical insights into this framework's ability to mitigate task-reward misalignment and present a practical implementation. Our experimental results show that our framework outperforms conventional IL baselines in complex and transfer learning scenarios.