Unlike short-form retrieval-augmented generation (RAG), such as factoid question answering, long-form RAG requires retrieval to provide documents covering a wide range of relevant information. Automated report generation exemplifies this setting: it requires not only relevant information but also a more elaborate response with comprehensive information. Yet, existing retrieval methods are primarily optimized for relevance ranking rather than information coverage. To address this limitation, we propose LANCER, an LLM-based reranking method for nugget coverage. LANCER predicts what sub-questions should be answered to satisfy an information need, predicts which documents answer these sub-questions, and reranks documents in order to provide a ranked list covering as many information nuggets as possible. Our empirical results show that LANCER enhances the quality of retrieval as measured by nugget coverage metrics, achieving higher $α$-nDCG and information coverage than other LLM-based reranking methods. Our oracle analysis further reveals that sub-question generation plays an essential role.

翻译：与短形式检索增强生成（例如事实性问答）不同，长形式检索增强生成要求检索提供的文档能覆盖广泛的相关信息。自动报告生成是这种场景的典型示例：它不仅需要相关信息，还需要生成包含全面信息的更详尽回答。然而，现有的检索方法主要针对相关性排序进行优化，而非信息覆盖。为应对这一局限，我们提出了LANCER，一种基于大语言模型、旨在实现信息单元覆盖的重排序方法。LANCER预测为满足信息需求应回答哪些子问题，预测哪些文档能回答这些子问题，并对文档进行重排序，以提供一个能覆盖尽可能多信息单元的排序列表。我们的实验结果表明，根据信息单元覆盖指标（如α-nDCG和信息覆盖率）衡量，LANCER提升了检索质量，其表现优于其他基于大语言模型的重排序方法。我们的理想情况分析进一步表明，子问题生成在其中起着至关重要的作用。