NBDI：一种用于从任务无关演示中提取技能的简单有效终止条件
NBDI: A Simple and Effective Termination Condition for Skill Extraction from Task-Agnostic Demonstrations
AI总结：
针对固定长度技能易错过决策点、限制长时序复杂任务学习效率的问题，研究提出基于新颖性的决策点识别方法NBDI，通过状态-动作新颖性模块利用经验数据确定技能终止点，性能优于基线且环境适应性强。
摘要：
智能体能够基于不同的粒度和时长水平做出决策。技能学习的最新进展使智能体能够通过有效引导其选择合适的技能来解决复杂的长时序任务。然而，使用固定长度技能的做法很容易导致错过有价值的决策点，这最终限制了进一步探索和更快策略学习的潜力。在本研究中，我们提出学习一种简单有效的终止条件，该条件通过一个利用智能体经验数据的状态-动作新颖性模块来识别决策点。我们的方法——基于新颖性的决策点识别（Novelty-based Decision Point Identification, NBDI），在复杂的长时序任务中表现优于以往的基线方法，即使在下游任务的环境配置存在显著变化的情况下依然有效，凸显了决策点识别在技能学习中的重要性。
英文摘要：
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.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
