← 首页|学术|Beyond Memory Majority: Latent-Source Reasoning for Knowledge Conflict
cs.AI · 2608.19701 · 2026-08-20

Beyond Memory Majority: Latent-Source Reasoning for Multi-Agent Memory Arbitration

Lin, Chenchen; Yuan, Wenhao; Wang, Xuehe; Ngai, Edith Cheuk Han
一句话:多 agent 记忆有共同偏差,投票失效——潜在来源推理修正仲裁

问题

多 agent 系统积累的记忆来自不同 agent,现有方法按投票/加权合并,但忽略了共同来源导致的相关偏差

方法

潜在来源推理:对检索到的记忆建模其来源相关性,在仲裁时纠正共享偏差

结果

在共享偏差场景下大幅优于独立证据合并基线;记忆来源溯源改善最终决策质量

与研究方向的关联

多智能体协作系统中记忆共享的核心问题,影响长期知识积累的可靠性

原文摘要

Long-term multi-agent systems continuously accumulate the memories produced by different agents. Existing memory methods typically treat retrieved memories as independent evidence and combine them through voting or weighting. However, this independence assumption often fails in multi-agent settings: memories written by different agents may inherit the same upstream source or shared bias, causing correlated evidence to be repeatedly counted and creating a false majority. We term this failure mode
Multi-AgentMemorySource BiasArbitration
ArXiv 2026-08-22 日报精读 · 返回简报