用几何距离优化检索排序，快3倍且精度不降。
Reranker Optimization via Geodesic Distances on k-NN Manifolds
- 基于k-近邻流形计算测地距离，融合全局与局部语义结构。
- 在8个数据集上优于基线，最快仅4.7毫秒/查询。
- 适合需要低延迟的实时检索增强生成场景。
当前检索增强生成（RAG）的神经重排序方法依赖交叉编码器或大语言模型，需大量计算资源，单次查询延迟达3-5秒。本文提出Maniscope，一种基于k-近邻流形上测地距离的几何重排序方法。该方法结合全局余弦相似度与局部流形几何，捕捉平面欧氏度量遗漏的语义结构。在8个BEIR基准数据集（1,233个查询）上，Maniscope在三个最难数据集上超越基于HNSW图的基线：NFCorpus提升7.0%、TREC-COVID提升1.6%、AorB提升2.8%（NDCG@3），且速度更快（平均4.7毫秒 vs 14.8毫秒）。相比交叉编码器重排序器，Maniscope实现98%以上准确率，延迟降低10-45倍。在TREC-COVID上，大语言模型重排序仅带来0.5%的额外提升，但延迟高出840倍，凸显Maniscope在实时RAG部署中的实用性。该方法复杂度为O(N D + M^2 D + M k log k)，其中M << N，支持亚10毫秒延迟。计划开源发布。
原文摘要 · Abstract (English)
Current neural reranking approaches for retrieval-augmented generation (RAG) rely on cross-encoders or large language models (LLMs), requiring substantial computational resources and exhibiting latencies of 3-5 seconds per query. We propose Maniscope, a geometric reranking method that computes geodesic distances on k-nearest neighbor (k-NN) manifolds constructed over retrieved document candidates. This approach combines global cosine similarity with local manifold geometry to capture semantic structure that flat Euclidean metrics miss. Evaluating on eight BEIR benchmark datasets (1,233 queries), Maniscope outperforms HNSW graph-based baseline on the three hardest datasets (NFCorpus: +7.0%, TREC-COVID: +1.6%, AorB: +2.8% NDCG@3) while being 3.2x faster (4.7 ms vs 14.8 ms average). Compared to cross-encoder rerankers, Maniscope achieves within 2% accuracy at 10-45x lower latency. On TREC-COVID, LLM-Reranker provides only +0.5% NDCG@3 improvement over Maniscope at 840x higher latency, positioning Maniscope as a practical alternative for real-time RAG deployment. The method requires O(N D + M^2 D + M k log k) complexity where M << N , enabling sub-10 ms latency. We plan to release Maniscope as open-source software.
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。