DocuSearch：结合知识图谱扩展与逐块接地评估的混合 RAG 企业文档搜索
原标题：Hybrid Retrieval-Augmented Generation with Knowledge Graph Expansion, RRF Fusion, and Per-Chunk Grounded Evaluation for Enterprise Document Search
AI 摘要
DocuSearch 是一个面向企业文档搜索的混合检索增强生成系统，在电信网络运维环境中开发并评估。它结合了基于 Qdrant 的语义搜索、基于 SQLite FTS5 的 BM25 全文搜索以及知识图谱邻居扩展，通过加权 RRF 融合，并采用交叉编码器重排序和 MMR 进行选择。其核心创新是逐块智能体评估循环，用于验证上下文充分性和答案接地性，未接地的答案会被拒绝。在内部电信文档语料库上，DocuSearch 相比仅密集检索的 RAG 基线，在 Precision@10、Recall@10 和接地率上均有显著提升。
正文节选
Hybrid Retrieval-Augmented Generation with Knowledge Graph Expansion, RRF Fusion, and Per-Chunk Grounded Evaluation for Enterprise Document Search Abstract Getting accurate, grounded answers out of large enterprise document repositories is a difficult problem. Dense vector retrieval alone frequently performs poorly on queries that mix technical terminology, vendor-specific acronyms, or require reasoning across several non-adjacent sections. DocuSearch was built to address exactly this gap — an o