arXiv 2606.045222026-06-03cs.IRcs.AIcs.DBcs.LG

ANN搜索：召回真正重要的

ANN Search: Recall What Matters

University of Ioannina（伊奥尼亚大学）；Archimedes, Athena RC（阿基米德，雅典RC）

Dimitris Dimitropoulos, Nikos Mamoulis

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出用逆近似比1/Ratio@k替代Recall@k来评估近似最近邻搜索质量，实验表明前者能更准确反映实际效用并降低计算开销。

摘要：

近似最近邻（ANN）搜索已成为信息检索和现代机器学习任务（从分类到检索增强生成）的核心原语。社区主要通过给定Recall@k（检索到的真实精确最近邻的比例）下的吞吐量来评估和调优ANN算法。我们认为，ANN搜索真正重要的是检索结果的质量，而非它们与真实kNN集合的重叠。我们证明，使用Recall@k评估检索质量会带来不必要的计算开销，并研究用逆近似比1/Ratio@k替代它。1/Ratio@k评估检索到的邻居与真实邻居之间距离的差异。它无需判断、无需超参数，仅通过标准ANN基准输入即可计算。我们在涵盖广泛内在维度的多样化数据集上对最先进的ANN算法进行基准测试，从效率、下游分类和检索增强生成三个维度全面评估这两个指标。在效率方面，优化1/Ratio@k达到操作质量阈值所需的计算成本远低于Recall@k。在下游任务中，即使Recall@k显著下降，性能指标（标签精度、语义相似度、BERTScore和LLM评分质量）仍保持高度稳定。相反，逆近似比紧密反映了这种稳定性，比Recall@k更好地追踪实际效用。最终，虽然Recall@k夸大了近似的真实成本，但1/Ratio@k提供了更准确、可部署的ANN实际质量代理。

英文摘要：

Approximate nearest neighbor (ANN) search has become a core primitive in information retrieval and modern machine learning tasks, from classification to retrieval-augmented generation. The community evaluates and tunes ANN algorithms primarily on their throughput at a given Recall@k, the fraction of true exact neighbors retrieved. We argue that what really matters in ANN search is the quality of the retrieved results and not their overlap with the true kNN set. We show that using Recall@k to assess retrieval quality forces unnecessary computational overhead and investigate replacing it by 1/Ratio@k, the inverse approximation ratio. 1/Ratio@k evaluates the differences between the distances of the retrieved and true neighbors. It is judge-free, hyperparameter-free, and computable from standard ANN benchmark inputs alone. We benchmark state-of-the-art ANN algorithms across diverse datasets spanning a wide range of intrinsic dimensionalities, evaluating the two metrics comprehensively across efficiency, downstream classification, and retrieval-augmented generation. On the efficiency axis, optimizing for 1/Ratio@k reaches operational quality thresholds at a substantially lower computational cost than Recall@k. In downstream tasks, performance indicators (label precision, semantic similarity, BERTScore, and LLM-graded quality) remain highly stable even when Recall@k drops significantly. The inverse approximation ratio, on the other hand, closely mirrors this stability, tracking true utility much better than Recall@k. Ultimately, while Recall@k overstates the true cost of approximation, 1/Ratio@k offers a more accurate, deployable proxy for actual ANN quality.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

AI总结：

摘要：

英文摘要：

相关论文