DemoSpeedup：通过熵引导的演示加速来加速视觉运动策略
DemoSpeedup: Accelerating Visuomotor Policies via Entropy-Guided Demonstration Acceleration
AI总结：
提出 DemoSpeedup，利用生成式策略估计逐帧动作熵并据此分段、变率下采样演示，使视觉运动策略最高提速3倍且成功率不降反升。
摘要：
模仿学习在机器人操作中展现出巨大前景，但由于人类操作者收集的演示通常较为迟缓，策略的执行往往慢得令人不满。在本工作中，我们提出 DemoSpeedup，这是一种通过熵引导的演示加速来提升视觉运动策略执行速度的自监督方法。DemoSpeedup 首先在常速演示上训练任意生成式策略（例如 ACT 或 Diffusion Policy），并将其用作逐帧动作熵估计器。其关键洞见是：动作熵估计较低的帧要求更一致的策略行为，这通常表明需要更高精度的操作；相比之下，熵估计较高的帧对应更随意的阶段，因此可以更安全地加速。于是，我们根据估计的熵对原始演示进行分段，并以随熵值增大而提高的速率进行下采样来加速这些演示。使用加速后的演示进行训练，所得策略的执行速度最高可达原来的 3 倍，同时保持任务完成性能。有趣的是，由于决策视野缩短带来的益处，这些策略甚至能够取得比使用常速演示训练的策略更高的成功率。项目页面：https://demospeedup.github.io/
英文摘要：
Imitation learning has shown great promise in robotic manipulation, but the policy's execution is often unsatisfactorily slow due to commonly tardy demonstrations collected by human operators. In this work, we present DemoSpeedup, a self-supervised method to accelerate visuomotor policy execution via entropy-guided demonstration acceleration. DemoSpeedup starts from training an arbitrary generative policy (e.g., ACT or Diffusion Policy) on normal-speed demonstrations, which serves as a per-frame action entropy estimator. The key insight is that frames with lower action entropy estimates call for more consistent policy behaviors, which often indicate the demands for higher-precision operations. In contrast, frames with higher entropy estimates correspond to more casual sections, and therefore can be more safely accelerated. Thus, we segment the original demonstrations according to the estimated entropy, and accelerate them by down-sampling at rates that increase with the entropy values. Trained with the speedup demonstrations, the resulting policies execute up to 3 times faster while maintaining the task completion performance. Interestingly, these policies could even achieve higher success rates than those trained with normal-speed demonstrations, due to the benefits of reduced decision-making horizons. Project Page: https://demospeedup.github.io/
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。
