提出分层记忆架构，在内存饱和时仍能保留关键信息并高效检索。
HTM-EAR: Importance-Preserving Tiered Memory with Hybrid Routing under Saturation
- 分两级存储：快速工作区与归档库，结合重要性评估与混合路由。
- 在15000条数据中保持查询精度（MRR=1.000），接近无限记忆表现。
- 适合长期运行的智能体，尤其对记忆持久性和准确性要求高场景。
长周期智能体面临内存约束，需在有限上下文内结构化管理积累的事实并保留关键信息。本文提出HTM-EAR，一种分层记忆架构，将基于HNSW的工作内存（L1，容量500）与归档存储（L2，容量5000）结合，采用重要性感知的淘汰策略与混合路由机制。当L1满时，按重要性与使用频率加权评分淘汰项；查询优先在L1处理，若相似度或实体覆盖率不足，则回退至L2，候选结果通过交叉编码器重排序。在持续饱和条件下（15,000条事实），基于合成流与真实BGL系统日志进行五次随机种子测试。消融实验对比了无重排序、无路由门、LRU淘汰及无限记忆的基线。在饱和下，完整系统保持活跃查询精度（MRR=1.000），实现可控遗忘，逼近基线表现（0.997 ± 0.003）。而LRU虽延迟低（21.1毫秒），但永久丢弃2416条关键事实。在BGL日志上，完整系统达MRR 0.336，接近基线（0.370），而LRU降至0.069。
原文摘要 · Abstract (English)
Memory constraints in long-running agents require structured management of accumulated facts while preserving essential information under bounded context limits. We introduce HTM-EAR, a hierarchical tiered memory substrate that integrates HNSW-based working memory (L1) with archival storage (L2), combining importance-aware eviction and hybrid routing. When L1 reaches capacity, items are evicted using a weighted score of importance and usage. Queries are first resolved in L1; if similarity or entity coverage is insufficient, retrieval falls back to L2, and candidates are re-ranked using a cross-encoder. We evaluate the system under sustained saturation (15,000 facts; L1 capacity 500; L2 capacity 5000) using synthetic streams across five random seeds and real BGL system logs. Ablation studies compare the full system against variants without cross-encoder re-ranking, without routing gates, with LRU eviction, and an oracle with unbounded memory. Under saturation, the full model preserves active-query precision (MRR = 1.000) while enabling controlled forgetting of stale history, approaching oracle active performance (0.997 +/- 0.003). In contrast, LRU minimizes latency (21.1 ms) but permanently evicts 2416 essential facts. On BGL logs, the full system achieves MRR 0.336, close to the oracle (0.370), while LRU drops to 0.069. Code is publicly available at: https://github.com/shubham-61291/HTM-EAR
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。