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/.

翻译：具身智能体面临一个根本性局限：一旦部署至真实环境，它们便难以获取新知识以提升任务表现。本文提出Dejavu，一种通用部署后学习框架，通过经验反馈网络（EFN）将检索到的执行记忆注入冻结的视觉-语言-动作（VLA）策略中。EFN能识别与上下文相关的先前动作经验，并基于检索到的引导信息对动作预测进行条件约束。我们采用强化学习与语义相似度奖励训练EFN，促使预测动作在当前观测下与历史行为对齐。在部署过程中，EFN持续用新轨迹扩展其记忆库，使智能体展现出"从经验中学习"的能力。跨多样具身任务的实验表明，EFN在适应性、鲁棒性及成功率上均优于冻结基线。项目页面：https://dejavu2025.github.io/。