用排序信号指导大模型重写查询，提升检索效果
GRAPE: Let GRPO Supervise Query Rewriting by Ranking for Retrieval
- 用分组相对策略优化融合排序信号，指导查询重写
- 多语言/长文本/多模态场景下召回率平均提升4.9%
- 无需重训检索器，适合现有系统快速升级
CLIP模型在大规模检索系统中扮演核心角色，但在多语言、长文本或跨模态查询下性能常下降。为避免重训检索器或重新嵌入语料库的高昂成本，我们提出GRAPE（分组排序感知策略优化增强），一种即插即用方法，利用基于大模型的查询重写弥补差距。不同于缺乏显式监督的现有方法，GRAPE通过分组相对策略优化（GRPO）将排序信号注入重写大模型，确保重写后的查询更贴合冻结检索器的潜在分布。关键发现：朴素的相似度微调会引发得分膨胀现象——无关候选获得过高评分，为此我们提出基于语料库相对排序的奖励机制加以缓解。在多语言（Flickr30k-CN、CVLUE、XM3600）、长文本（Wikipedia）和多模态（CIRR）基准上广泛实验表明，GRAPE持续提升性能，召回率@10平均提升4.9%，且无需修改底层检索器。代码已开源。
原文摘要 · Abstract (English)
The CLIP model has established itself as a cornerstone of large-scale retrieval systems. However, its performance often degrades under distributional shifts such as multilingual, long-form, or multimodal queries. To avoid the prohibitive costs associated with retriever retraining or corpus re-embedding, we propose GRAPE (Grouped Ranking-Aware Policy Optimization Enhancement), a plug-and-play approach that leverages LLM-based query rewriting to bridge these gaps. Unlike existing methods that lack explicit supervision, GRAPE integrates ranking signals into the rewriting LLM via Grouped Relative Policy Optimization (GRPO), ensuring rewritten queries are better aligned with the frozen retriever's latent distribution. Crucially, we identify a score inflation phenomenon in naive similarity-based finetuning - where irrelevant candidates receive indiscriminately high scores - and mitigate it with a novel corpus-relative ranking-based reward. Extensive experiments across multilingual (Flickr30k-CN, CVLUE, XM3600), long-form (Wikipedia), and multimodal (CIRR) benchmarks demonstrate that GRAPE consistently improves performance, achieving an average gain of 4.9% in Recall@10 without any modification to the underlying retriever. The code is available at https://github.com/mogulzhang/GRAPE.
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。