@inproceedings{zhan-etal-2023-similarizing,
title = "Similarizing the Influence of Words with Contrastive Learning to Defend Word-level Adversarial Text Attack",
author = "Zhan, Pengwei and
Yang, Jing and
Wang, He and
Zheng, Chao and
Huang, Xiao and
Wang, Liming",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2023",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-acl.500",
doi = "10.18653/v1/2023.findings-acl.500",
pages = "7891--7906",
abstract = "Neural language models are vulnerable to word-level adversarial text attacks, which generate adversarial examples by directly substituting discrete input words. Previous search methods for word-level attacks assume that the information in the important words is more influential on prediction than unimportant words. In this paper, motivated by this assumption, we propose a self-supervised regularization method for Similarizing the Influence of Words with Contrastive Learning (SIWCon) that encourages the model to learn sentence representations in which words of varying importance have a more uniform influence on prediction. Experiments show that SIWCon is compatible with various training methods and effectively improves model robustness against various unforeseen adversarial attacks. The effectiveness of SIWCon is also intuitively shown through qualitative analysis and visualization of the loss landscape, sentence representation, and changes in model confidence.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhan-etal-2023-similarizing">
<titleInfo>
<title>Similarizing the Influence of Words with Contrastive Learning to Defend Word-level Adversarial Text Attack</title>
</titleInfo>
<name type="personal">
<namePart type="given">Pengwei</namePart>
<namePart type="family">Zhan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jing</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">He</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chao</namePart>
<namePart type="family">Zheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiao</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Liming</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Rogers</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordan</namePart>
<namePart type="family">Boyd-Graber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naoaki</namePart>
<namePart type="family">Okazaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Toronto, Canada</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Neural language models are vulnerable to word-level adversarial text attacks, which generate adversarial examples by directly substituting discrete input words. Previous search methods for word-level attacks assume that the information in the important words is more influential on prediction than unimportant words. In this paper, motivated by this assumption, we propose a self-supervised regularization method for Similarizing the Influence of Words with Contrastive Learning (SIWCon) that encourages the model to learn sentence representations in which words of varying importance have a more uniform influence on prediction. Experiments show that SIWCon is compatible with various training methods and effectively improves model robustness against various unforeseen adversarial attacks. The effectiveness of SIWCon is also intuitively shown through qualitative analysis and visualization of the loss landscape, sentence representation, and changes in model confidence.</abstract>
<identifier type="citekey">zhan-etal-2023-similarizing</identifier>
<identifier type="doi">10.18653/v1/2023.findings-acl.500</identifier>
<location>
<url>https://aclanthology.org/2023.findings-acl.500</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>7891</start>
<end>7906</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Similarizing the Influence of Words with Contrastive Learning to Defend Word-level Adversarial Text Attack
%A Zhan, Pengwei
%A Yang, Jing
%A Wang, He
%A Zheng, Chao
%A Huang, Xiao
%A Wang, Liming
%Y Rogers, Anna
%Y Boyd-Graber, Jordan
%Y Okazaki, Naoaki
%S Findings of the Association for Computational Linguistics: ACL 2023
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F zhan-etal-2023-similarizing
%X Neural language models are vulnerable to word-level adversarial text attacks, which generate adversarial examples by directly substituting discrete input words. Previous search methods for word-level attacks assume that the information in the important words is more influential on prediction than unimportant words. In this paper, motivated by this assumption, we propose a self-supervised regularization method for Similarizing the Influence of Words with Contrastive Learning (SIWCon) that encourages the model to learn sentence representations in which words of varying importance have a more uniform influence on prediction. Experiments show that SIWCon is compatible with various training methods and effectively improves model robustness against various unforeseen adversarial attacks. The effectiveness of SIWCon is also intuitively shown through qualitative analysis and visualization of the loss landscape, sentence representation, and changes in model confidence.
%R 10.18653/v1/2023.findings-acl.500
%U https://aclanthology.org/2023.findings-acl.500
%U https://doi.org/10.18653/v1/2023.findings-acl.500
%P 7891-7906
Markdown (Informal)
[Similarizing the Influence of Words with Contrastive Learning to Defend Word-level Adversarial Text Attack](https://aclanthology.org/2023.findings-acl.500) (Zhan et al., Findings 2023)
ACL