让大模型理解文本结构，生成更连贯的问答与摘要。
Disco-RAG: Discourse-Aware Retrieval-Augmented Generation
- 用话语树和修辞图建模段落内与跨段落结构
- 在无需微调下超越现有方法，在问答与长文摘要上表现最优
- 适合需要深度理解文本逻辑的智能问答与文档处理场景
检索增强生成（RAG）已成为提升大语言模型在知识密集型任务中性能的重要手段。然而，现有大多数RAG策略将检索到的段落以扁平无结构的方式处理，难以捕捉文本结构线索，限制了模型从分散证据中综合知识的能力。为此，我们提出Disco-RAG，一种注重话语结构的框架，通过显式注入话语信号来增强生成过程。该方法构建段落内的话语树以捕捉局部层级关系，并建立段落间的修辞图以建模跨段落连贯性。这些结构被联合整合为一个规划蓝图，用于指导生成。在问答与长文档摘要基准上的实验表明，该方法在无需微调的情况下达到当前最优性能。结果凸显了话语结构在推进RAG系统中的关键作用。
原文摘要 · Abstract (English)
Retrieval-Augmented Generation (RAG) has emerged as an important means of enhancing the performance of large language models (LLMs) in knowledge-intensive tasks. However, most existing RAG strategies treat retrieved passages in a flat and unstructured way, which prevents the model from capturing structural cues and constrains its ability to synthesize knowledge from dispersed evidence across documents. To overcome these limitations, we propose Disco-RAG, a discourse-aware framework that explicitly injects discourse signals into the generation process. Our method constructs intra-chunk discourse trees to capture local hierarchies and builds inter-chunk rhetorical graphs to model cross-passage coherence. These structures are jointly integrated into a planning blueprint that conditions the generation. Experiments on question answering and long-document summarization benchmarks show the efficacy of our approach. Disco-RAG achieves state-of-the-art results on the benchmarks without fine-tuning. These findings underscore the important role of discourse structure in advancing 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。