NBDI: A Simple and Effective Termination Condition for Skill Extraction from Task-Agnostic Demonstrations
This work addresses a bottleneck in skill learning for intelligent agents by improving decision point identification, though it appears incremental as it builds on existing skill learning methods.
The paper tackled the problem of fixed-length skills limiting exploration and policy learning in complex tasks by proposing a termination condition called NBDI that identifies decision points using a state-action novelty module, resulting in outperforming previous baselines in long-horizon tasks and maintaining effectiveness across varied environments.
Intelligent agents are able to make decisions based on different levels of granularity and duration. Recent advances in skill learning enabled the agent to solve complex, long-horizon tasks by effectively guiding the agent in choosing appropriate skills. However, the practice of using fixed-length skills can easily result in skipping valuable decision points, which ultimately limits the potential for further exploration and faster policy learning. In this work, we propose to learn a simple and effective termination condition that identifies decision points through a state-action novelty module that leverages agent experience data. Our approach, Novelty-based Decision Point Identification (NBDI), outperforms previous baselines in complex, long-horizon tasks, and remains effective even in the presence of significant variations in the environment configurations of downstream tasks, highlighting the importance of decision point identification in skill learning.