门控记忆策略

中文摘要 AI 辅助

机器人操作任务表现出不同的记忆需求，从需要无记忆的马尔可夫任务到依赖历史信息的非马尔可夫任务。简单扩展视觉-运动政策的观察历史会导致性能下降，因为分布偏移和过拟合。为此，我们提出门控记忆策略（GMP），一种视觉-运动策略，能够学习何时回溯记忆和回溯什么。GMP采用学习的记忆门机制，仅在必要时激活历史上下文，提高鲁棒性和反应性。为高效学习回溯内容，GMP引入轻量级交叉注意力模块，构建有效的潜在记忆表示。为进一步增强鲁棒性，GMP向历史动作注入扩散噪声，在训练和推理过程中减轻对噪声或不准确历史的敏感性。在我们提出的非马尔可夫基准MemMimic上，GMP在长历史基线上的平均成功率提高了30.1%，同时在RoboMimic上的马尔可夫任务中保持竞争力。所有代码、数据和真实场景部署说明均可在项目网站https://gated-memory-policy.github.io/上获得。

英文摘要

Robotic manipulation tasks exhibit varying memory requirements, ranging from Markovian tasks that require no memory to non-Markovian tasks that demand in-context memorization of historical information within a single trial or in-context adaptation based on the outcomes of multiple past trials. Surprisingly, simply extending observation histories of a visuomotor policy often leads to a significant performance drop due to distribution shift and overfitting. To address these issues, we propose Gated Memory Policy (GMP), a visuomotor policy that learns both when to recall memory and what to recall. To learn when to recall memory, GMP employs a learned memory gate mechanism that selectively activates history context only when necessary, improving robustness and reactivity. To learn what to recall efficiently, GMP introduces a lightweight cross-attention module that constructs effective latent memory representations. To further enhance robustness, GMP injects diffusion noise into historical actions, mitigating sensitivity to noisy or inaccurate histories during both training and inference. On our proposed non-Markovian benchmark MemMimic, GMP achieves a 30.1% average success rate improvement over long-history baselines, while maintaining competitive performance on Markovian tasks in RoboMimic. All code, data and in-the-wild deployment instructions are available on our project website https://gated-memory-policy.github.io/.

发表机构

• Stanford University（斯坦福大学）

机构由 AI 辅助整理，请以论文原文为准。