@inproceedings{yang-etal-2023-enough,
title = "Once is Enough: A Light-Weight Cross-Attention for Fast Sentence Pair Modeling",
author = "Yang, Yuanhang and
Qi, Shiyi and
Liu, Chuanyi and
Wang, Qifan and
Gao, Cuiyun and
Xu, Zenglin",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.emnlp-main.168/",
doi = "10.18653/v1/2023.emnlp-main.168",
pages = "2800--2806",
abstract = "Transformer-based models have achieved great success on sentence pair modeling tasks, such as answer selection and natural language inference (NLI). These models generally perform cross-attention over input pairs, leading to prohibitive computational cost. Recent studies propose dual-encoder and late interaction architectures for faster computation. However, the balance between the expressive of cross-attention and computation speedup still needs better coordinated. To this end, this paper introduces a novel paradigm TopicAns for efficient sentence pair modeling. TopicAns involves a lightweight cross-attention mechanism. It conducts query encoding only once while modeling the query-candidate interaction in parallel. Extensive experiments conducted on four tasks demonstrate that our TopicAnscan speed up sentence pairing by over 113x while achieving comparable performance as the more expensive cross-attention models."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yang-etal-2023-enough">
<titleInfo>
<title>Once is Enough: A Light-Weight Cross-Attention for Fast Sentence Pair Modeling</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yuanhang</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shiyi</namePart>
<namePart type="family">Qi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chuanyi</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qifan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Cuiyun</namePart>
<namePart type="family">Gao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zenglin</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Houda</namePart>
<namePart type="family">Bouamor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Juan</namePart>
<namePart type="family">Pino</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kalika</namePart>
<namePart type="family">Bali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Singapore</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Transformer-based models have achieved great success on sentence pair modeling tasks, such as answer selection and natural language inference (NLI). These models generally perform cross-attention over input pairs, leading to prohibitive computational cost. Recent studies propose dual-encoder and late interaction architectures for faster computation. However, the balance between the expressive of cross-attention and computation speedup still needs better coordinated. To this end, this paper introduces a novel paradigm TopicAns for efficient sentence pair modeling. TopicAns involves a lightweight cross-attention mechanism. It conducts query encoding only once while modeling the query-candidate interaction in parallel. Extensive experiments conducted on four tasks demonstrate that our TopicAnscan speed up sentence pairing by over 113x while achieving comparable performance as the more expensive cross-attention models.</abstract>
<identifier type="citekey">yang-etal-2023-enough</identifier>
<identifier type="doi">10.18653/v1/2023.emnlp-main.168</identifier>
<location>
<url>https://aclanthology.org/2023.emnlp-main.168/</url>
</location>
<part>
<date>2023-12</date>
<extent unit="page">
<start>2800</start>
<end>2806</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Once is Enough: A Light-Weight Cross-Attention for Fast Sentence Pair Modeling
%A Yang, Yuanhang
%A Qi, Shiyi
%A Liu, Chuanyi
%A Wang, Qifan
%A Gao, Cuiyun
%A Xu, Zenglin
%Y Bouamor, Houda
%Y Pino, Juan
%Y Bali, Kalika
%S Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F yang-etal-2023-enough
%X Transformer-based models have achieved great success on sentence pair modeling tasks, such as answer selection and natural language inference (NLI). These models generally perform cross-attention over input pairs, leading to prohibitive computational cost. Recent studies propose dual-encoder and late interaction architectures for faster computation. However, the balance between the expressive of cross-attention and computation speedup still needs better coordinated. To this end, this paper introduces a novel paradigm TopicAns for efficient sentence pair modeling. TopicAns involves a lightweight cross-attention mechanism. It conducts query encoding only once while modeling the query-candidate interaction in parallel. Extensive experiments conducted on four tasks demonstrate that our TopicAnscan speed up sentence pairing by over 113x while achieving comparable performance as the more expensive cross-attention models.
%R 10.18653/v1/2023.emnlp-main.168
%U https://aclanthology.org/2023.emnlp-main.168/
%U https://doi.org/10.18653/v1/2023.emnlp-main.168
%P 2800-2806
Markdown (Informal)
[Once is Enough: A Light-Weight Cross-Attention for Fast Sentence Pair Modeling](https://aclanthology.org/2023.emnlp-main.168/) (Yang et al., EMNLP 2023)
ACL