arXiv Analytics

Sign in

arXiv:2009.07258 [cs.IR]AbstractReferencesReviewsResources

BERT-QE: Contextualized Query Expansion for Document Re-ranking

Zhi Zheng, Kai Hui, Ben He, Xianpei Han, Le Sun, Andrew Yates

Published 2020-09-15Version 1

Query expansion aims to mitigate the mismatch between the language used in a query and in a document. Query expansion methods can suffer from introducing non-relevant information when expanding the query, however. To bridge this gap, inspired by recent advances in applying contextualized models like BERT to the document retrieval task, this paper proposes a novel query expansion model that leverages the strength of the BERT model to better select relevant information for expansion. In evaluations on the standard TREC Robust04 and GOV2 test collections, the proposed BERT-QE model significantly outperforms BERT-Large models commonly used for document retrieval.

Related articles:
arXiv:1904.07094 [cs.IR] (Published 2019-04-15)
Contextualized Word Representations for Document Re-Ranking