用信息增益剔除冗余证据，提升问答生成质量与效率
Less is More for RAG: Information Gain Pruning for Generator-Aligned Reranking and Evidence Selection
- 根据生成器反馈选择有用证据，动态过滤无效内容
- 多数据集测试中F1提升12%-20%，输入 tokens 减少76%-79%
- 无需修改接口，适配现有RAG系统，部署友好
检索增强生成（RAG）通过外部证据为大模型提供支持，但在有限上下文预算下，关键挑战在于决定哪些检索到的段落应被注入。我们发现，传统相关性指标（如NDCG）与端到端问答质量相关性较弱，甚至在多段落注入时出现负相关，因冗余和轻微冲突会扰乱生成。为此，我们提出信息增益剪枝（IGP），一个部署友好的重排序与剪枝模块，利用生成器对齐的效用信号选择证据，并在截断前过滤弱或有害段落，无需更改现有预算接口。在五个开放域问答基准上，使用多种检索器与生成器，IGP均持续改善质量-成本权衡。在一个典型多证据场景中，相比仅检索器基线，IGP实现约12%-20%的相对F1提升，同时将最终输入令牌数减少约76%-79%。
原文摘要 · Abstract (English)
Retrieval-augmented generation (RAG) grounds large language models with external evidence, but under a limited context budget, the key challenge is deciding which retrieved passages should be injected. We show that retrieval relevance metrics (e.g., NDCG) correlate weakly with end-to-end QA quality and can even become negatively correlated under multi-passage injection, where redundancy and mild conflicts destabilize generation. We propose \textbf{Information Gain Pruning (IGP)}, a deployment-friendly reranking-and-pruning module that selects evidence using a generator-aligned utility signal and filters weak or harmful passages before truncation, without changing existing budget interfaces. Across five open-domain QA benchmarks and multiple retrievers and generators, IGP consistently improves the quality--cost trade-off. In a representative multi-evidence setting, IGP delivers about +12--20% relative improvement in average F1 while reducing final-stage input tokens by roughly 76--79% compared to retriever-only baselines.
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。