IBM发布STAIR检索增强生成论文
Great RAG paper from IBM. There are some really good ideas on how to solve common RAG issues. It's...
IBM这篇论文用目录结构解决RAG分块问题，在检索准确率和降低幻觉方面都有显著提升。
IBM提出STAIR方法，利用目录结构解决RAG检索器分块丢弃文档层次的问题。该研究在SearchTome基准上达到82.6%的Recall @1，显著优于76.9%的微调DSI模型。STAIR将目录作为生成式检索器的寻址方案，幻觉率低于0.05%，且在少量训练样本下表现良好。
Great RAG paper from IBM. There are some really good ideas on how to solve common RAG issues. It's...
Great RAG paper from IBM. There are some really good ideas on how to solve common RAG issues. It's well known that retrievers chunk long documents by length, which discards the hierarchy the document already has. So they propose using a table of contents. A table of contents helps to encodes exactly the global structure that chunking throws away. STAIR uses that table of contents as the addressing scheme for a generative retriever, so the model stores and retrieves information from its own parameters against a structure the corpus supplies. On SearchTome, it reaches Recall @1 of 82.6 percent against 76.9 percent for a fine-tuned Differentiable Search Index, a statistically significant gap, with BM25 at 59.5 percent and DPR at 68.7 percent. Hallucination stays below 0.05 percent, which is the standing objection to generative retrieval and the reason grounding the address space in a real hierarchy is worth the extra structure. The ablations also show it generalizes where very few training samples exist. Paper: academy.dair.ai/papers/stair-s… 💬 4 🔄 0 ❤️ 9 👀 1357 📊 6 ⚡
