← 研踪 PaperTrailCVF · CVPR 2026:cvf:c76e581b68bc8ed0536d460b141412f1 · 记录 #1
Dejavu: Towards Experience Feedback Learning for Embodied Intelligence
Wu, ShaokaiJi, YanbiaoLi, QiuchangZhang, ZhiyiHe, QichenXie, WenyuanZhang, GuodongBayramli, BayramDing, YueLu, Hongtao 来源日期（仅到年份） 2026
来源记录日期 2026
雷达观测 2026-11-01 22:58
摘要
Embodied agents face a fundamental limitation: once deployed in real-world environments, they cannot easily acquire new knowledge to improve task performance. In this paper, we propose Dejavu, a general post-deployment learning framework that augments a frozen Vision-Language-Action (VLA) policy with retrieved execution memories through an Experience Feedback Network (EFN). EFN identifies contextually relevant prior action experiences and conditions action prediction on the retrieved guidance. We train EFN with reinforcement learning and semantic similarity rewards, encouraging the predicted actions to align with past behaviors under the current observation. During deployment, EFN continually expands its memory with new trajectories, enabling the agent to exhibit "learning from experience." Experiments across diverse embodied tasks show that EFN improves adaptability, robustness, and success rates over frozen baselines. Our Project Page is https://dejavu2025.github.io/.
论文发现出处
这些页面提供论文线索，不代表同行评审或质量认证。
观测到的版本
平台记录号代表本站观测到的元数据变化，不是作者修订号。来源日期仅到月份或年份时，按该月或该年首日存档。
记录 #1 · 2026 · 最新观测版本
Dejavu: Towards Experience Feedback Learning for Embodied Intelligence
Embodied agents face a fundamental limitation: once deployed in real-world environments, they cannot easily acquire new knowledge to improve task performance. In this paper, we propose Dejavu, a general post-deployment learning framework that augments a frozen Vision-Language-Action (VLA) policy with retrieved execution memories through an Experience Feedback Network (EFN). EFN identifies contextually relevant prior action experiences and conditions action prediction on the retrieved guidance. We train EFN with reinforcement learning and semantic similarity rewards, encouraging the predicted actions to align with past behaviors under the current observation. During deployment, EFN continually expands its memory with new trajectories, enabling the agent to exhibit "learning from experience." Experiments across diverse embodied tasks show that EFN improves adaptability, robustness, and success rates over frozen baselines. Our Project Page is https://dejavu2025.github.io/.
查看此版本解读 →查看此版本原文 ↗