Rewind-IL：在线故障检测与状态重置用于模仿学习
Rewind-IL: Online Failure Detection and State Respawning for Imitation Learning
AI总结：
Rewind-IL通过零样本故障检测和状态重置机制，提升模仿学习在长时域任务中的可靠性，解决执行漂移问题。
摘要：
模仿学习使机器人能够从演示中获取复杂的视觉-运动操作技能，但部署失败仍是主要障碍，尤其针对长时域动作分块策略。一旦执行偏离演示流，这些策略常继续产生局部合理动作而无法恢复。现有运行监控要么需要故障数据，要么在良性特征漂移下误触发，或仅停留在故障检测而不提供恢复机制。我们提出Rewind-IL，一种无需训练的在线保护框架，用于生成式动作分块模仿策略。Rewind-IL结合基于时间间块差异估计（TIDE）的零样本故障检测器，经分割符合预测校准，与一种返回机器人到语义验证的安全中间状态的状态重置机制。离线时，视觉-语言模型在演示中识别恢复检查点，冻结的策略编码器用于构建紧凑的检查点特征数据库。在线时，Rewind-IL监控重叠动作块的一致性，跟踪与检查点库的相似性，并在故障时将执行回最新验证的安全状态，然后从干净的策略状态重新启动推理。在真实世界和模拟长时域操作任务中的实验，包括转移到流匹配动作分块策略，证明了策略内部一致性与语义基础的重置相结合，为改进模仿学习的可靠性提供了实用途径。补充材料可在https://sjay05.github.io/rewind-il获取。
英文摘要：
Imitation learning has enabled robots to acquire complex visuomotor manipulation skills from demonstrations, but deployment failures remain a major obstacle, especially for long-horizon action-chunked policies. Once execution drifts off the demonstration manifold, these policies often continue producing locally plausible actions without recovering from the failure. Existing runtime monitors either require failure data, over-trigger under benign feature drift, or stop at failure detection without providing a recovery mechanism. We present Rewind-IL, a training-free online safeguard framework for generative action-chunked imitation policies. Rewind-IL combines a zero-shot failure detector based on Temporal Inter-chunk Discrepancy Estimate (TIDE), calibrated with split conformal prediction, with a state-respawning mechanism that returns the robot to a semantically verified safe intermediate state. Offline, a vision-language model identifies recovery checkpoints in demonstrations, and the frozen policy encoder is used to construct a compact checkpoint feature database. Online, Rewind-IL monitors self-consistency in overlapping action chunks, tracks similarity to the checkpoint library, and, upon failure, rewinds execution to the latest verified safe state before restarting inference from a clean policy state. Experiments on real-world and simulated long-horizon manipulation tasks, including transfer to flow-matching action-chunked policies, demonstrate that policy-internal consistency coupled with semantically grounded respawning offers a practical route to improved reliability in imitation learning. Supplemental materials are available at https://sjay05.github.io/rewind-il
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。
