No Stupid Questions: An Analysis of Question Query Generation for Citation Recommendation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913887091163136 |
|---|---|
| author | Zimmerman, Brian D. Aubert-Béduchaud, Julien Boudin, Florian Aizawa, Akiko Vechtomova, Olga |
| author_facet | Zimmerman, Brian D. Aubert-Béduchaud, Julien Boudin, Florian Aizawa, Akiko Vechtomova, Olga |
| contents | Existing techniques for citation recommendation are constrained by their adherence to article contents and metadata. We leverage GPT-4o-mini's latent expertise as an inquisitive assistant by instructing it to ask questions which, when answered, could expose new insights about an excerpt from a scientific article. We evaluate the utility of these questions as retrieval queries, measuring their effectiveness in retrieving and ranking masked target documents. In some cases, generated questions ended up being better queries than extractive keyword queries generated by the same model. We additionally propose MMR-RBO, a variation of Maximal Marginal Relevance (MMR) using Rank-Biased Overlap (RBO) to identify which questions will perform competitively with the keyword baseline. As all question queries yield unique result sets, we contend that there are no stupid questions. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_08196 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | No Stupid Questions: An Analysis of Question Query Generation for Citation Recommendation Zimmerman, Brian D. Aubert-Béduchaud, Julien Boudin, Florian Aizawa, Akiko Vechtomova, Olga Information Retrieval Digital Libraries Existing techniques for citation recommendation are constrained by their adherence to article contents and metadata. We leverage GPT-4o-mini's latent expertise as an inquisitive assistant by instructing it to ask questions which, when answered, could expose new insights about an excerpt from a scientific article. We evaluate the utility of these questions as retrieval queries, measuring their effectiveness in retrieving and ranking masked target documents. In some cases, generated questions ended up being better queries than extractive keyword queries generated by the same model. We additionally propose MMR-RBO, a variation of Maximal Marginal Relevance (MMR) using Rank-Biased Overlap (RBO) to identify which questions will perform competitively with the keyword baseline. As all question queries yield unique result sets, we contend that there are no stupid questions. |
| title | No Stupid Questions: An Analysis of Question Query Generation for Citation Recommendation |
| topic | Information Retrieval Digital Libraries |
| url | https://arxiv.org/abs/2506.08196 |