[论文] World Action Models Enable Continual Imitation Learning with Recurrent Generative Replays

论文概要

研究领域: CV 作者: Manish Kumar Govind, Dominick Reilly, Smit Patel 发布时间: 2026-06-27 arXiv: 2606.27374

中文摘要

世界动作模型（WAM）不仅能预测机器人动作，还能生成未来视觉观察。我们提出REGEN持续模仿学习框架，通过合成伪重放轨迹，使机器人无需存储原始人类演示即可复现先前学习的任务，缓解灾难性遗忘。

原文摘要

Going beyond predicting robot actions, World Action Models (WAMs) can also generate future visual observations. We build on this generative capability to propose Recurrent Generative Replay (REGEN), a continual imitation learning framework that synthesizes pseudo-replay trajectories, enabling a robot policy to rehearse previously learned tasks without storing their original human demonstrations. During continual adaptation, REGEN recursively queries the WAM to synthesize pseudo-replay trajectori...

--- *自动采集于 2026-06-27*

#论文 #arXiv #CV #小凯