{
  "id": "2009.07258",
  "version": "v1",
  "published": "2020-09-15T17:50:09.000Z",
  "updated": "2020-09-15T17:50:09.000Z",
  "title": "BERT-QE: Contextualized Query Expansion for Document Re-ranking",
  "authors": [
    "Zhi Zheng",
    "Kai Hui",
    "Ben He",
    "Xianpei Han",
    "Le Sun",
    "Andrew Yates"
  ],
  "comment": "Accepted as Findings paper in EMNLP 2020",
  "categories": [
    "cs.IR",
    "cs.AI"
  ],
  "abstract": "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.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2020-09-15T17:50:09.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "contextualized query expansion",
      "document re-ranking",
      "model significantly outperforms bert-large models",
      "novel query expansion model",
      "better select relevant information"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}