BMRetriever: Tuning Large Language Models as Better Biomedical Text Retrievers
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914963841351680 |
|---|---|
| author | Xu, Ran Shi, Wenqi Yu, Yue Zhuang, Yuchen Zhu, Yanqiao Wang, May D. Ho, Joyce C. Zhang, Chao Yang, Carl |
| author_facet | Xu, Ran Shi, Wenqi Yu, Yue Zhuang, Yuchen Zhu, Yanqiao Wang, May D. Ho, Joyce C. Zhang, Chao Yang, Carl |
| contents | Developing effective biomedical retrieval models is important for excelling at knowledge-intensive biomedical tasks but still challenging due to the deficiency of sufficient publicly annotated biomedical data and computational resources. We present BMRetriever, a series of dense retrievers for enhancing biomedical retrieval via unsupervised pre-training on large biomedical corpora, followed by instruction fine-tuning on a combination of labeled datasets and synthetic pairs. Experiments on 5 biomedical tasks across 11 datasets verify BMRetriever's efficacy on various biomedical applications. BMRetriever also exhibits strong parameter efficiency, with the 410M variant outperforming baselines up to 11.7 times larger, and the 2B variant matching the performance of models with over 5B parameters. The training data and model checkpoints are released at \url{https://huggingface.co/BMRetriever} to ensure transparency, reproducibility, and application to new domains. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_18443 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | BMRetriever: Tuning Large Language Models as Better Biomedical Text Retrievers Xu, Ran Shi, Wenqi Yu, Yue Zhuang, Yuchen Zhu, Yanqiao Wang, May D. Ho, Joyce C. Zhang, Chao Yang, Carl Computation and Language Artificial Intelligence Information Retrieval Quantitative Methods Developing effective biomedical retrieval models is important for excelling at knowledge-intensive biomedical tasks but still challenging due to the deficiency of sufficient publicly annotated biomedical data and computational resources. We present BMRetriever, a series of dense retrievers for enhancing biomedical retrieval via unsupervised pre-training on large biomedical corpora, followed by instruction fine-tuning on a combination of labeled datasets and synthetic pairs. Experiments on 5 biomedical tasks across 11 datasets verify BMRetriever's efficacy on various biomedical applications. BMRetriever also exhibits strong parameter efficiency, with the 410M variant outperforming baselines up to 11.7 times larger, and the 2B variant matching the performance of models with over 5B parameters. The training data and model checkpoints are released at \url{https://huggingface.co/BMRetriever} to ensure transparency, reproducibility, and application to new domains. |
| title | BMRetriever: Tuning Large Language Models as Better Biomedical Text Retrievers |
| topic | Computation and Language Artificial Intelligence Information Retrieval Quantitative Methods |
| url | https://arxiv.org/abs/2404.18443 |