外观
论文arXiv 2609.226841 个月前
StateMem：用于视觉-语言-动作策略的单状态残差记忆与自适应推理
原文链接：StateMem: Single-State Residual Memory with Adaptive Inference for Vision-Language-Action Policies（arXiv 2609.22684 · 发布日期：2026-09-19）
StateMem: Single-State Residual Memory with Adaptive Inference for Vision-Language-Action Policies
StateMem通过单状态残差记忆和自适应前缀路由提升VLA历史信息保留，并在LIBERO、RoboMemArena及真机操作中验证。
核心要点
- 依赖记忆的操作需利用早期交互信息，而现有VLA多依赖当前观测。
- StateMem用预测误差经低秩残差更新持久记忆token，并自适应路由缓存前缀。
- 免训练控制器在线调整路由阈值，快速校正补偿缓存复用中的过时前缀特征。
- 在LIBERO达97.6%平均成功率；RoboMemArena遮挡类单VLA最佳，真机六任务平均成功率提升21%。
验证范围与局限：验证范围包括LIBERO、RoboMemArena和六个真实世界操作任务。
英文摘要 (Abstract)
Memory-dependent robotic manipulation often requires later actions to use information from earlier interactions. Existing vision-language-action (VLA) policies primarily rely on current observations, limiting historical information retention. Memory-augmented VLAs, such as MemoryVLA, address this limitation with external memory banks but require explicit storage and retrieval. To address these limitations, we propose StateMem, a single-state residual memory framework for VLA policies that uses prediction error to update a persistent memory token through low-rank residuals and to adaptively route cached prefixes. A training-free controller adjusts the routing threshold online, while fast correction compensates for stale prefix features during cache reuse. We evaluate StateMem on LIBERO, RoboMemArena, and real-world manipulation tasks. On LIBERO, StateMem achieves an average success rate of 97.6% and reduces the average VLM prefix refresh rate by 20.25% relative to full refresh. In the Occlusion category of RoboMemArena, StateMem achieves the best performance among single-VLA methods, reaching 21.8% Task Success Rate (TSR) and 44.3% Cumulative Success Rate (CSR). Across six real-world manipulation tasks, it achieves +21% in average success rate.
作者：Wenzhuo Li, Qiongfeng Shi, Yi Zhou
来源
本页事实对应的公开页面，以原文为准。