arXiv:2510.10181v3 Announce Type: replace-cross
Abstract: 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/.
Adaptation to free-living drives loss of beneficial endosymbiosis through metabolic trade-offs
Symbioses are widespread (1) and underpin the function of diverse ecosystems (2-6), but their evolutionary stability is challenging to explain (7,8). Fitness trade-offs between con-trasting