CEQE: Contextualized Embeddings for Query Expansion

Naseri, S., Dalton, J. , Yates, A. and Allan, J. (2021) CEQE: Contextualized Embeddings for Query Expansion. In: 43rd European Conference on IR Research, ECIR 2021, 28 Mar - 1 Apr 2021, ISBN 9783030721121 (doi:10.1007/978-3-030-72113-8_31)

[img] Text
237968.pdf - Accepted Version

297kB

Abstract

In this work we leverage recent advances in context-sensitive language models to improve the task of query expansion. Contextualized word representation models, such as ELMo and BERT, are rapidly replacing static embedding models. We propose a new model, Contextualized Embeddings for Query Expansion (CEQE), that utilizes query-focused contextualized embedding vectors. We study the behavior of contextual representations generated for query expansion in ad-hoc document retrieval. We conduct our experiments on probabilistic retrieval models as well as in combination with neural ranking models. We evaluate CEQE on two standard TREC collections: Robust and Deep Learning. We find that CEQE outperforms static embedding-based expansion methods on multiple collections (by up to 18% on Robust and 31% on Deep Learning on average precision) and also improves over proven probabilistic pseudo-relevance feedback (PRF) models. We further find that multiple passes of expansion and reranking result in continued gains in effectiveness with CEQE-based approaches outperforming other approaches. The final model incorporating neural and CEQE-based expansion score achieves gains of up to 5% in P@20 and 2% in AP on Robust over the state-of-the-art transformer-based re-ranking model, Birch.

Item Type:Conference Proceedings
Status:Published
Refereed:Yes
Glasgow Author(s) Enlighten ID:Yates, Professor Andrew and Dalton, Dr Jeff
Authors: Naseri, S., Dalton, J., Yates, A., and Allan, J.
College/School:College of Science and Engineering > School of Computing Science
ISBN:9783030721121
Copyright Holders:Copyright © 2021 Springer International Publishing
First Published:First published in Advances in Information Retrieval Part I, LCNS 12656
Publisher Policy:Reproduced in accordance with the publisher copyright policy

University Staff: Request a correction | Enlighten Editors: Update this record