根据对话类型自动选择最佳检索路径，提升长时记忆召回率。
SelRoute: Query-Type-Aware Routing for Long-Term Conversational Memory Retrieval
- 按查询类型分配不同检索管道：词法、语义、混合或词汇增强
- 在LongMemEval_M上达到0.800的Recall@5，优于现有方法
- 无需GPU和大模型推理，适合资源受限场景
从长期对话记忆中检索相关历史交互通常依赖大型密集检索模型（110M-1.5B参数）或大语言模型增强索引。本文提出SelRoute框架，根据查询类型将请求路由至专用检索管道——词法、语义、混合或词汇增强。在LongMemEval_M数据集上，使用bge-base-en-v1.5（109M参数）时，SelRoute取得0.800的Recall@5；使用bge-small-en-v1.5（33M参数）时为0.786，均高于使用LLM生成事实键的Contriever（0.762）。仅用SQLite FTS5的零模型基线也达NDCG@5=0.692，超过所有已有基线，部分归因于词法检索实现差异。五折分层交叉验证显示路由稳定（CV差距1.3-2.4点Recall@5），4/6查询类型在各折中保持一致。基于正则表达式的查询类型分类器准确率达83%，端到端预测类型检索仍优于统一基线（Recall@5=0.689）。跨基准测试覆盖8个额外数据集，含62,000+实例（如MSDialog、LoCoMo、QReCC、PerLTQA），未调参即表现良好，但在推理密集型任务中暴露短板（RECOR Recall@5=0.149）。还发现词汇扩展在存储时提升词法检索但降低嵌入检索，提示应按管道分别决策。系统全程无需GPU与大模型推理。
原文摘要 · Abstract (English)
Retrieving relevant past interactions from long-term conversational memory typically relies on large dense retrieval models (110M-1.5B parameters) or LLM-augmented indexing. We introduce SelRoute, a framework that routes each query to a specialized retrieval pipeline -- lexical, semantic, hybrid, or vocabulary-enriched -- based on its query type. On LongMemEval_M (Wu et al., 2024), SelRoute achieves Recall@5 of 0.800 with bge-base-en-v1.5 (109M parameters) and 0.786 with bge-small-en-v1.5 (33M parameters), compared to 0.762 for Contriever with LLM-generated fact keys. A zero-ML baseline using SQLite FTS5 alone achieves NDCG@5 of 0.692, already exceeding all published baselines on ranking quality -- a gap we attribute partly to implementation differences in lexical retrieval. Five-fold stratified cross-validation confirms routing stability (CV gap of 1.3-2.4 Recall@5 points; routes stable for 4/6 query types across folds). A regex-based query-type classifier achieves 83% effective routing accuracy, and end-to-end retrieval with predicted types (Recall@5 = 0.689) still outperforms uniform baselines. Cross-benchmark evaluation on 8 additional benchmarks spanning 62,000+ instances -- including MSDialog, LoCoMo, QReCC, and PerLTQA -- confirms generalization without benchmark-specific tuning, while exposing a clear failure mode on reasoning-intensive retrieval (RECOR Recall@5 = 0.149) that bounds the claim. We also identify an enrichment-embedding asymmetry: vocabulary expansion at storage time improves lexical search but degrades embedding search, motivating per-pipeline enrichment decisions. The full system requires no GPU and no LLM inference at query time.
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。