AI 中文总结
研究针对大语言模型智能体长期记忆中多跳关联未被充分测量的问题，提出了MemHop基准和ProGraph两层内存架构，通过概要扩展与压缩残差实现跨实体遍历，提升了多跳推理及精确召回能力，在多个基准测试中表现优异并开源相关实现。
AI 中文摘要
长期记忆对于跨会话交互的大语言模型智能体至关重要，但当前内存基准主要评估单跳召回，多跳关联基本未被测量。本文有三点贡献。一是引入MemHop，这是一个跨10个社交网络场景、跳深度为1至5的1000个问题的多跳内存基准，并带有每跳证据注释。二是提出概要图内存（ProGraph），这是一种两层内存架构，结合了概要扩展（通过在大语言模型编写的概要叙事中自然出现的实体名称进行子串匹配遍历）和压缩残差（与每次概要更新共同提取的确切日期、数量和命名项，无需额外API成本）。三是全网格消融显示了跨基准机制的专业化：概要扩展驱动多跳推理（移除时在MemHop上降低22.6个百分点），压缩残差驱动精确召回（未共同提取时在LoCoMo上降低8.6个百分点），且在单一架构内交叉效应低于3个百分点。ProGraph在MemHop上平均为80.1%（与FullContext参考匹配），在LoCoMo上为78.4%（比FullContext高出11.3个百分点），在两者上均优于Mem0、A-Mem、HippoRAG和RAG。我们发布了MemHop、ProGraph和基线实现。
英文摘要
Long-term memory is essential for LLM agents that interact across sessions, yet current memory benchmarks primarily evaluate single-hop recall, leaving multi-hop association largely unmeasured. We make three contributions. First, we introduce MemHop, a multi-hop memory benchmark of 1,000 questions at hop depths 1-5 across 10 social-network scenarios, with per-hop evidence annotations. Second, we present Profile-Graph Memory (ProGraph), a two-layer memory architecture combining (i) profile expansion -- substring-matched traversal of entity names that naturally appear in LLM-written profile narratives, a minimal alternative to explicit knowledge-graph construction -- and (ii) compression residuals -- exact dates, quantities, and named items co-extracted with each profile update at zero extra API cost. Third, a full-grid ablation shows cross-benchmark mechanism specialization: profile expansion drives multi-hop reasoning (-22.6pp on MemHop when removed) while compression residuals drive precision recall (-8.6pp on LoCoMo when not co-extracted), with cross-effects under 3pp within a single architecture. ProGraph averages 80.1% on MemHop (matching the FullContext reference) and 78.4% on LoCoMo (exceeding FullContext by 11.3pp), outperforming Mem0, A-Mem, HippoRAG, and RAG on both. We release MemHop, ProGraph, and baseline implementations.
Comments11 pages, 2 figures, 7 tables. Code and MemHop benchmark: https://github.com/ShengtongZhu/ProGraph
