Sparse, Dense, and Attentional Representations for Text Retrieval

Yi Luan, Jacob Eisenstein, Kristina Toutanova, Michael Collins


Abstract
Dual encoders perform retrieval by encoding documents and queries into dense low-dimensional vectors, scoring each document by its inner product with the query. We investigate the capacity of this architecture relative to sparse bag-of-words models and attentional neural networks. Using both theoretical and empirical analysis, we establish connections between the encoding dimension, the margin between gold and lower-ranked documents, and the document length, suggesting limitations in the capacity of fixed-length encodings to support precise retrieval of long documents. Building on these insights, we propose a simple neural model that combines the efficiency of dual encoders with some of the expressiveness of more costly attentional architectures, and explore sparse-dense hybrids to capitalize on the precision of sparse retrieval. These models outperform strong alternatives in large-scale retrieval.
Anthology ID:
2021.tacl-1.20
Volume:
Transactions of the Association for Computational Linguistics, Volume 9
Month:
Year:
2021
Address:
Cambridge, MA
Editors:
Brian Roark, Ani Nenkova
Venue:
TACL
SIG:
Publisher:
MIT Press
Note:
Pages:
329–345
Language:
URL:
https://aclanthology.org/2021.tacl-1.20
DOI:
10.1162/tacl_a_00369
Bibkey:
Cite (ACL):
Yi Luan, Jacob Eisenstein, Kristina Toutanova, and Michael Collins. 2021. Sparse, Dense, and Attentional Representations for Text Retrieval. Transactions of the Association for Computational Linguistics, 9:329–345.
Cite (Informal):
Sparse, Dense, and Attentional Representations for Text Retrieval (Luan et al., TACL 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.tacl-1.20.pdf
Video:
 https://aclanthology.org/2021.tacl-1.20.mp4