Dynamic Injection of Entity Knowledge into Dense Retrievers

Ikuya Yamada, Ryokan Ri, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo


Abstract
Dense retrievers often struggle with queries involving less-frequent entities due to their limited entity knowledge. We propose the Knowledgeable Passage Retriever (KPR), a BERT-based retriever enhanced with a context-entity attention layer and dynamically updatable entity embeddings. This design enables KPR to incorporate external entity knowledge without retraining. Experiments on three datasets demonstrate that KPR consistently improves retrieval accuracy, with particularly large gains on the EntityQuestions dataset. When built on the off-the-shelf bge-base retriever, KPR achieves state-of-the-art performance among similarly sized models on two datasets. Models and code are released at https://github.com/knowledgeable-embedding/knowledgeable-embedding.
Anthology ID:
2025.findings-emnlp.915
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2025
Month:
November
Year:
2025
Address:
Suzhou, China
Editors:
Christos Christodoulopoulos, Tanmoy Chakraborty, Carolyn Rose, Violet Peng
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
16867–16879
Language:
URL:
https://aclanthology.org/2025.findings-emnlp.915/
DOI:
Bibkey:
Cite (ACL):
Ikuya Yamada, Ryokan Ri, Takeshi Kojima, Yusuke Iwasawa, and Yutaka Matsuo. 2025. Dynamic Injection of Entity Knowledge into Dense Retrievers. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 16867–16879, Suzhou, China. Association for Computational Linguistics.
Cite (Informal):
Dynamic Injection of Entity Knowledge into Dense Retrievers (Yamada et al., Findings 2025)
Copy Citation:
PDF:
https://aclanthology.org/2025.findings-emnlp.915.pdf
Checklist:
 2025.findings-emnlp.915.checklist.pdf