LANCER: LLM Reranking for Nugget Coverage
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914291734544384 |
|---|---|
| author | Ju, Jia-Huei Landry, François G. Yang, Eugene Verberne, Suzan Yates, Andrew |
| author_facet | Ju, Jia-Huei Landry, François G. Yang, Eugene Verberne, Suzan Yates, Andrew |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_22008 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | LANCER: LLM Reranking for Nugget Coverage Ju, Jia-Huei Landry, François G. Yang, Eugene Verberne, Suzan Yates, Andrew Information Retrieval 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. |
| title | LANCER: LLM Reranking for Nugget Coverage |
| topic | Information Retrieval |
| url | https://arxiv.org/abs/2601.22008 |