Explain Like I am BM25: Interpreting a Dense Model’s Ranked-List with a Sparse Approximation

Llordes, M., Ganguly, D. , Bhatia, S. and Agarwal, C. (2023) Explain Like I am BM25: Interpreting a Dense Model’s Ranked-List with a Sparse Approximation. In: 46th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR23), Taipei, Taiwan, 23-27 July 2023, pp. 1976-1980. ISBN 9781450394086 (doi: 10.1145/3539618.3591982)

Abstract

Neural retrieval models (NRMs) have been shown to outperform their statistical counterparts owing to their ability to capture semantic meaning via dense document representations. These models, however, suffer from poor interpretability as they do not rely on explicit term matching. As a form of local per-query explanations, we introduce the notion of equivalent queries that are generated by maximizing the similarity between the NRM's results and the result set of a sparse retrieval system with the equivalent query. We then compare this approach with existing methods such as RM3-based query expansion and contrast differences in retrieval effectiveness and in the terms generated by each approach.

• Item Type: Conference Proceedings
• Keywords: Interpretability, explainability, neural ranking models.
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Ganguly, Dr Debasis and Llordes Richards, Mr Michael Jon
• Authors: Llordes, M., Ganguly, D., Bhatia, S., and Agarwal, C.
• College/School: College of Science and Engineering > School of Computing Science
• ISBN: 9781450394086
• Copyright Holders: Copyright © 2023 held by the owner/author(s)
• First Published: First published in SIGIR '23: Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval
• Publisher Policy: Reproduced in accordance with the publisher copyright policy