arXiv：cs.AI· Yilin Wu, Rutav Shah, Shuaiyi Huang, Sumedh A. Sontakke, Homanga Bharadhwaj, Andrea Bajcsy, Max Simchowitz, Guanya Shi·· 18 小时前
IMPRINT：为机器人策略学习多流潜在记忆，无需筛选历史
Memory Without Curation: Learning Multi-Stream Latent Memory for Robot Policies
AI 导读
研究者提出 IMPRINT，一种策略无关的潜在记忆框架，为 VLA 模型加入多流记忆且不显式约束策略该记住什么。它在推理时增量更新潜在记忆以降低延迟，并采用两阶段训练：先监督记忆重建观测、动作、本体感受等多流信息并估计任务进度，再与策略端到端联合微调。在仿真和真实记忆任务上，IMPRINT 平均成功率比基础 VLA 至少提升 26%，并保持记忆增强 VLA 中最低的推理延迟。
正文
Abstract:It is widely appreciated that robotic policies require long-term memory to perform complex tasks. However, directly conditioning policies on all past observations leaves policy learning susceptible to learning spurious shortcuts, and introduces large latency at policy deployment. As such, today's memory-augmented policies curate what to remember, either by retaining only recent frames, selected keyframes, or textual summaries, or by learning latent representations of the visual stream alone. In this work, we propose an alternative approach IMPRINT, a policy-agnostic latent memory framework that augments vision-language-action (VLA) models with multi-stream memory without explicitly constraining what the policy should remember. IMPRINT compresses different streams into a latent memory that is updated incrementally at inference, enabling low latency. To avoid shortcut learning, IMPRINT adopts a two-stage training procedure, first supervising memory to reconstruct multiple streams (observations, actions, proprioception) of past information, as well as estimate current task progress, followed by end-to-end finetuning, jointly with the policy, on action prediction. Across both simulation and real-world tasks demanding memory, IMPRINT improves average success over the base VLA by at least 26%, outperforming popular alternatives, and maintaining lowest inference latency among memory-augmented VLAs. Through careful ablations, we demonstrate the utility of incorporating diverse information streams into robotic memory, validating the advantages of foregoing history curation. Project website: this https URL.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.14406 [cs.RO] |
| (or arXiv:2610.14406v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.14406 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yilin Wu [view email]
[v1]
Fri, 9 Oct 2026 23:58:57 UTC (34,563 KB)
来源：arXiv：cs.AI · arxiv.org