提升图谱-RAG的多跳问答能力，通过结构化提示与上下文压缩解决推理瓶颈。
The Reasoning Bottleneck in Graph-RAG: Structured Prompting and Context Compression for Multi-Hop QA
- 用SPARQL链式思维提示将问题拆解为实体-关系三元组查询。
- 结合图遍历压缩上下文，减少60%信息量且不依赖大模型调用。
- 小模型经增强后性能媲美甚至超越大模型，成本降低12倍。
Graph-RAG系统通过将文档索引至知识图谱实现强大的多跳问答能力，但强检索不等于强回答。在三个多跳问答基准（HotpotQA、MuSiQue、2WikiMultiHopQA）上评估领先系统KET-RAG发现：77%至91%的问题其正确答案存在于检索到的上下文中，但准确率仅为35%至78%，其中73%至84%的错误源于推理失败。为此提出两项改进：(i) SPARQL链式思维提示（SPARQL CoT），将问题分解为与实体-关系上下文对齐的三元组查询；(ii) 图遍历压缩（graph-walk compression），通过知识图谱遍历将上下文压缩约60%，且无需调用大语言模型。SPARQL CoT使准确率提升2至14个百分点；图遍历压缩在小型模型上与结构化提示结合时平均提升6个百分点。令人意外的是，结合问题类型路由策略，一个完全增强的开源轻量级Llama-8B模型在所有三个基准上达到或超过未增强的Llama-70B基线表现，成本仅为其1/12。在LightRAG上的复现验证了该方法在不同Graph-RAG系统间的可迁移性。
原文摘要 · Abstract (English)
Graph-RAG systems achieve strong multi-hop question answering by indexing documents into knowledge graphs, but strong retrieval does not guarantee strong answers. Evaluating KET-RAG, a leading Graph-RAG system, on three multi-hop QA benchmarks (HotpotQA, MuSiQue, 2WikiMultiHopQA), we find that 77% to 91% of questions have the gold answer in the retrieved context, yet accuracy is only 35% to 78%, and 73% to 84% of errors are reasoning failures. We propose two augmentations: (i) SPARQL chain-of-thought prompting, which decomposes questions into triple-pattern queries aligned with the entity-relationship context, and (ii) graph-walk compression, which compresses the context by ~60% via knowledge-graph traversal with no LLM calls. SPARQL CoT improves accuracy by +2 to +14 pp; graph-walk compression adds +6 pp on average when paired with structured prompting on smaller models. Surprisingly, we show that, with question-type routing, a fully augmented budget open-weight Llama-8B model matches or exceeds the unaugmented Llama-70B baseline on all three benchmarks at ~12x lower cost. A replication on LightRAG confirms that our augmentations transfer across Graph-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。