首次系统分析查询重写如何影响RAG检索偏见，发现重写可减缓54%偏见但不全能。
Masking or Mitigating? Deconstructing the Impact of Query Rewriting on Retriever Biases in RAG
- 对比五种重写方法，发现大模型重写整体降偏效果最佳
- 简单重写靠提升得分方差减偏，伪文档生成则通过解耦特征实现
- 不同检索器对重写响应差异大，需按具体偏见选策略
在检索增强生成（RAG）系统中，密集检索器存在显著偏见，包括简短性、位置、字面匹配和重复偏见，可能损害检索质量。尽管查询重写已成为RAG标准流程，其对这些偏见的影响仍未知。本文首次系统研究查询增强技术对密集检索偏见的影响，评估了五种方法在六种检索器上的表现。结果表明，基于大模型的简单重写在整体上实现最强偏见缓解（54%），但在多种偏见共现的对抗条件下失效。机制分析揭示两种不同路径：简单重写通过增加得分方差降低偏见，而伪文档生成方法通过与偏见诱导特征真正解耦实现减偏。然而，无一种技术能统一解决所有偏见，且效果在不同检索器间差异显著。研究为根据特定偏见漏洞选择查询增强策略提供了实用指导。更广泛地，我们提出一个分类框架，区分查询-文档交互偏见与文档编码偏见，明确了仅从查询侧干预在去偏中的局限性。
原文摘要 · Abstract (English)
Dense retrievers in retrieval-augmented generation (RAG) systems exhibit systematic biases -- including brevity, position, literal matching, and repetition biases -- that can compromise retrieval quality. Query rewriting techniques are now standard in RAG pipelines, yet their impact on these biases remains unexplored. We present the first systematic study of how query enhancement techniques affect dense retrieval biases, evaluating five methods across six retrievers. Our findings reveal that simple LLM-based rewriting achieves the strongest aggregate bias reduction (54\%), yet fails under adversarial conditions where multiple biases combine. Mechanistic analysis uncovers two distinct mechanisms: simple rewriting reduces bias through increased score variance, while pseudo-document generation methods achieve reduction through genuine decorrelation from bias-inducing features. However, no technique uniformly addresses all biases, and effects vary substantially across retrievers. Our results provide practical guidance for selecting query enhancement strategies based on specific bias vulnerabilities. More broadly, we establish a taxonomy distinguishing query-document interaction biases from document encoding biases, clarifying the limits of query-side interventions for debiasing RAG systems.
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。