KAIST 提出 WikiTrace:解决智能体记忆导致的任务退化问题
如果你在给自己的智能体做记忆系统,这篇得看看。KAIST 发现存错的教训会让智能体把本来能做的任务搞砸,他们的 WikiTrace 把每条经验链接到原始运行轨迹,出错前能回溯验证,五个基准上平均多拿 10.7 分。
KAIST 研究人员发现,智能体的记忆可能让原本能完成的任务失败,他们把这一现象命名为 memory-induced degradation(记忆诱导退化)。问题的根源在于记忆的存储方式:现有系统只从标注为成功或失败的运行记录中提炼经验教训,保存时丢弃了原始运行轨迹,一旦某条经验是错的,智能体没有任何途径去验证它。
研究团队提出的 WikiTrace 将每条经验存储为一条简短洞见,并附带三样东西:该次运行成败原因的摘要、原始运行轨迹、以及关联链接。当某条洞见模糊不清、与其他洞见冲突、或来自一次失败的运行时,智能体会在执行前通过终端工具沿着链接回溯查证。
在 ALFWorld、ScienceWorld、PlanCraft、AppWorld 和 tau3-bench retail 五个基准上,WikiTrace 平均比无记忆智能体高 10.7 分,比仅使用相同洞见的方案高 3.2 分;原本可解任务的失败率从 6.5% 降到 4.5%。对正在给智能体搭建记忆系统的开发者来说,这组数字说明记忆的验证机制比记忆本身更关键。
Interesting insights about agent memory.
Might be worth a read if you are building memory systems for your agents.
Can an agent's memory make it fail tasks it would solve without memory?
KAIST researchers find that it can, and they call it memory-induced degradation.
The cause is how lessons are stored. They infer lessons from runs labeled only as success or failure, then save them without the run itself. When a lesson is wrong, the agent has no way to verify it.
The authors introduce WikiTrace, which stores each lesson as a short insight linked to a summary of why the run succeeded or failed and to the raw trajectory. When an insight is vague, conflicts with another, or comes from a failed run, the agent follows the links with terminal tools before acting.
Across ALFWorld, ScienceWorld, PlanCraft, AppWorld, and tau3-bench retail, it scores 10.7 points above the agent without memory on average and 3.2 points above the same insights alone. The failure rate for previously solvable tasks drops from 6.5% to 4.5%.
https://t.co/3wU1elq35o
Chat with Paper: https://t.co/b182EkAvLh
本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:WikiTrace、KAIST、智能体、记忆系统,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明:本文原始线索采编自「DAIR.AI」(@dair_ai),经 AITOP 多维评分管线结构化提炼。