@inproceedings{shahgir-etal-2024-asymmetric,
title = "Asymmetric Bias in Text-to-Image Generation with Adversarial Attacks",
author = "Shahgir, Haz and
Kong, Xianghao and
Ver Steeg, Greg and
Dong, Yue",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.344",
doi = "10.18653/v1/2024.findings-acl.344",
pages = "5779--5796",
abstract = "The widespread use of Text-to-Image (T2I) models in content generation requires careful examination of their safety, including their robustness to adversarial attacks. Despite extensive research on adversarial attacks, the reasons for their effectiveness remain underexplored. This paper presents an empirical study on adversarial attacks against T2I models, focusing on analyzing factors associated with attack success rates (ASR). We introduce a new attack objective - entity swapping using adversarial suffixes and two gradient-based attack algorithms. Human and automatic evaluations reveal the asymmetric nature of ASRs on entity swap: for example, it is easier to replace {``}human{''} with {``}robot{''} in the prompt {``}a human dancing in the rain.{''} with an adversarial suffix, but the reverse replacement is significantly harder. We further propose probing metrics to establish indicative signals from the model{'}s beliefs to the adversarial ASR. We identify conditions that result in a success probability of 60{\%} for adversarial attacks and others where this likelihood drops below 5{\%}. The code and data are available at https://github.com/Patchwork53/AsymmetricAttack",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="shahgir-etal-2024-asymmetric">
<titleInfo>
<title>Asymmetric Bias in Text-to-Image Generation with Adversarial Attacks</title>
</titleInfo>
<name type="personal">
<namePart type="given">Haz</namePart>
<namePart type="family">Shahgir</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xianghao</namePart>
<namePart type="family">Kong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Greg</namePart>
<namePart type="family">Ver Steeg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Dong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-08</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2024</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lun-Wei</namePart>
<namePart type="family">Ku</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andre</namePart>
<namePart type="family">Martins</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vivek</namePart>
<namePart type="family">Srikumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Bangkok, Thailand</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>The widespread use of Text-to-Image (T2I) models in content generation requires careful examination of their safety, including their robustness to adversarial attacks. Despite extensive research on adversarial attacks, the reasons for their effectiveness remain underexplored. This paper presents an empirical study on adversarial attacks against T2I models, focusing on analyzing factors associated with attack success rates (ASR). We introduce a new attack objective - entity swapping using adversarial suffixes and two gradient-based attack algorithms. Human and automatic evaluations reveal the asymmetric nature of ASRs on entity swap: for example, it is easier to replace “human” with “robot” in the prompt “a human dancing in the rain.” with an adversarial suffix, but the reverse replacement is significantly harder. We further propose probing metrics to establish indicative signals from the model’s beliefs to the adversarial ASR. We identify conditions that result in a success probability of 60% for adversarial attacks and others where this likelihood drops below 5%. The code and data are available at https://github.com/Patchwork53/AsymmetricAttack</abstract>
<identifier type="citekey">shahgir-etal-2024-asymmetric</identifier>
<identifier type="doi">10.18653/v1/2024.findings-acl.344</identifier>
<location>
<url>https://aclanthology.org/2024.findings-acl.344</url>
</location>
<part>
<date>2024-08</date>
<extent unit="page">
<start>5779</start>
<end>5796</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Asymmetric Bias in Text-to-Image Generation with Adversarial Attacks
%A Shahgir, Haz
%A Kong, Xianghao
%A Ver Steeg, Greg
%A Dong, Yue
%Y Ku, Lun-Wei
%Y Martins, Andre
%Y Srikumar, Vivek
%S Findings of the Association for Computational Linguistics: ACL 2024
%D 2024
%8 August
%I Association for Computational Linguistics
%C Bangkok, Thailand
%F shahgir-etal-2024-asymmetric
%X The widespread use of Text-to-Image (T2I) models in content generation requires careful examination of their safety, including their robustness to adversarial attacks. Despite extensive research on adversarial attacks, the reasons for their effectiveness remain underexplored. This paper presents an empirical study on adversarial attacks against T2I models, focusing on analyzing factors associated with attack success rates (ASR). We introduce a new attack objective - entity swapping using adversarial suffixes and two gradient-based attack algorithms. Human and automatic evaluations reveal the asymmetric nature of ASRs on entity swap: for example, it is easier to replace “human” with “robot” in the prompt “a human dancing in the rain.” with an adversarial suffix, but the reverse replacement is significantly harder. We further propose probing metrics to establish indicative signals from the model’s beliefs to the adversarial ASR. We identify conditions that result in a success probability of 60% for adversarial attacks and others where this likelihood drops below 5%. The code and data are available at https://github.com/Patchwork53/AsymmetricAttack
%R 10.18653/v1/2024.findings-acl.344
%U https://aclanthology.org/2024.findings-acl.344
%U https://doi.org/10.18653/v1/2024.findings-acl.344
%P 5779-5796
Markdown (Informal)
[Asymmetric Bias in Text-to-Image Generation with Adversarial Attacks](https://aclanthology.org/2024.findings-acl.344) (Shahgir et al., Findings 2024)
ACL