先筛选可访问记忆再检索，提升智能体记忆查询效率与准确率。
ShardMemo: Scope-Before-Routing for Agentic Memory Retrieval
- 先用元数据筛选允许访问的记忆块，再用学习模型精炼检索
- 在LoCoMo上比基线提升约3 F1，端到端评估最高提升6.8 F1
- 适合需要严格权限控制与高效检索的多任务智能体系统
智能体系统在跨会话、跨工具和跨任务中积累持久记忆，后续请求需在两个约束下检索：可访问范围与有限搜索预算。现有系统处理访问范围的方式存在缺陷：先检索后过滤会浪费探查预算在不合法记忆上；将范围视为学习特征则使硬性限制依赖路由质量。我们提出SHARD-MEMO，基于‘范围优先路由’设计：元数据谓词首先识别可访问的记忆分片，学习型路由器再从中选出少数分片进行局部近似最近邻检索。该方法将硬性可访问性与学习相关性分离，确保非法分片不消耗探查预算。在LoCoMo、HotpotQA和ToolBench上评估，覆盖对话、长上下文与程序化记忆场景。在匹配监督与固定预算条件下，相比学习型路由基线，SHARD-MEMO在LoCoMo上提升约+3 F1；在端到端LoCoMo评估中，相较最强基线最高提升+6.8 F1，HotpotQA与ToolBench也均有增益。
原文摘要 · Abstract (English)
Agentic systems accumulate persistent memory across sessions, tools, and tasks, and a later request must retrieve from it under two distinct constraints: which memories it is permitted to access, and which are relevant under a limited search budget. Existing memory systems handle access scope in two flawed ways: applying scope after retrieval wastes probe budget on inadmissible memories, while treating scope as a learned ranking feature makes a hard constraint depend on router quality. We present SHARD-MEMO, an agentic memory system built on scope-before-routing: metadata predicates first identify the admissible shards, and a learned router then selects a small number of them for shard-local approximate nearest neighbor retrieval. Given the supplied scope predicate and metadata, this separates hard admissibility from learned relevance ranking, so inadmissible shards cannot consume shard-probe budget. We evaluate on LoCoMo, HotpotQA, and ToolBench, covering conversational, long-context, and procedural memory. Under matched supervision and fixed budgets, SHARDMEMO improves over a learned router baseline by roughly +3 F1 on LoCoMo; in end-to-end LoCoMo evaluation it improves over the strongest evaluated memory baseline by up to +6.8 F1, with gains on HotpotQA and ToolBench.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。