Dual Skipping Guidance for Document Retrieval with Learned Sparse Representations

Yifan Qiao; Yingrui Yang; Haixin Lin; Tianbo Xiong; Xiyue Wang; Tao Yang

Dual Skipping Guidance for Document Retrieval with Learned Sparse Representations

Information Retrieval 2022-04-26 v1

Authors: Yifan Qiao , Yingrui Yang , Haixin Lin , Tianbo Xiong , Xiyue Wang , Tao Yang

Abstract

This paper proposes a dual skipping guidance scheme with hybrid scoring to accelerate document retrieval that uses learned sparse representations while still delivering a good relevance. This scheme uses both lexical BM25 and learned neural term weights to bound and compose the rank score of a candidate document separately for skipping and final ranking, and maintains two top-k thresholds during inverted index traversal. This paper evaluates time efficiency and ranking relevance of the proposed scheme in searching MS MARCO TREC datasets.

Keywords

information retrieval sparse learning image retrieval

Cite

@article{arxiv.2204.11154,
  title  = {Dual Skipping Guidance for Document Retrieval with Learned Sparse Representations},
  author = {Yifan Qiao and Yingrui Yang and Haixin Lin and Tianbo Xiong and Xiyue Wang and Tao Yang},
  journal= {arXiv preprint arXiv:2204.11154},
  year   = {2022}
}

Related papers

View all related →