机器人一回头就失忆，这论文给它装记忆
机器人干活时经常遇到一个尴尬：眼前看到的画面一模一样，但该做的动作却不一样——比如拧螺丝，拧到第几圈了，光看画面根本看不出来，得靠“刚才发生了什么”。现有机器人策略大多没这个记忆，一回头就失忆。这篇论文先造了个基准测试 HIDE,15 个任务专门考机器人的“记忆”:数重复次数、回忆之前的状态、跟踪进度，而且故意设计成“画面相同、动作不同”的陷阱。测下来，现有策略确实大面积翻车。然后他们提出 SEEK 框架，给机器人装三种互补的记忆机制：留证据、记状态、跟进度。结果单个机制有的任务有用、有的反而拖后腿，三个合起来平均成功率最高，仿真和真机都验证了。这提醒我们：机器人要可靠，光靠“看”不够，得在心里记一本账。
📄 原文摘要(英文)
Recent advances in robot learning have enabled manipulation policies to perform increasingly diverse tasks and generalize across environments. However, reliable execution often depends on hidden task states that cannot be determined from current observations alone, making interaction history essential. We introduce HIDE, a benchmark for evaluating manipulation memory under partial observability. HIDE comprises 15 tasks covering repetition counting, historical-state recall, and execution-progress tracking, with randomized initial configurations and decision points where similar observations require different actions depending on prior events. We further propose SEEK, a framework combining three complementary memory mechanisms to retain historical evidence and track execution state. Evaluations reveal substantial limitations in existing policies on HIDE, while memory augmentation improves task success in both simulation and real-world experiments. Individual mechanisms benefit some tasks but can degrade others; their combination achieves the highest average success rate on HIDE among the evaluated configurations. These findings highlight the importance of maintaining internal representations of hidden task states and matching memory design to task-specific information requirements.