@inproceedings{du-etal-2024-uor,
title = "{UOR}: Universal Backdoor Attacks on Pre-trained Language Models",
author = "Du, Wei and
Li, Peixuan and
Zhao, Haodong and
Ju, Tianjie and
Ren, Ge and
Liu, Gongshen",
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.468",
doi = "10.18653/v1/2024.findings-acl.468",
pages = "7865--7877",
abstract = "Task-agnostic and transferable backdoors implanted in pre-trained language models (PLMs) pose a severe security threat as they can be inherited to any downstream task. However, existing methods rely on manual selection of triggers and backdoor representations, hindering their effectiveness and universality across different PLMs or usage paradigms. In this paper, we propose a new backdoor attack method called UOR, which overcomes these limitations by turning manual selection into automatic optimization. Specifically, we design poisoned supervised contrastive learning, which can automatically learn more uniform and universal backdoor representations. This allows for more even coverage of the output space, thus hitting more labels in downstream tasks after fine-tuning. Furthermore, we utilize gradient search to select appropriate trigger words that can be adapted to different PLMs and vocabularies. Experiments show that UOR achieves better attack performance on various text classification tasks compared to manual methods. Moreover, we test on PLMs with different architectures, usage paradigms, and more challenging tasks, achieving higher scores for universality.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="du-etal-2024-uor">
<titleInfo>
<title>UOR: Universal Backdoor Attacks on Pre-trained Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wei</namePart>
<namePart type="family">Du</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peixuan</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haodong</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tianjie</namePart>
<namePart type="family">Ju</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ge</namePart>
<namePart type="family">Ren</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gongshen</namePart>
<namePart type="family">Liu</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>Task-agnostic and transferable backdoors implanted in pre-trained language models (PLMs) pose a severe security threat as they can be inherited to any downstream task. However, existing methods rely on manual selection of triggers and backdoor representations, hindering their effectiveness and universality across different PLMs or usage paradigms. In this paper, we propose a new backdoor attack method called UOR, which overcomes these limitations by turning manual selection into automatic optimization. Specifically, we design poisoned supervised contrastive learning, which can automatically learn more uniform and universal backdoor representations. This allows for more even coverage of the output space, thus hitting more labels in downstream tasks after fine-tuning. Furthermore, we utilize gradient search to select appropriate trigger words that can be adapted to different PLMs and vocabularies. Experiments show that UOR achieves better attack performance on various text classification tasks compared to manual methods. Moreover, we test on PLMs with different architectures, usage paradigms, and more challenging tasks, achieving higher scores for universality.</abstract>
<identifier type="citekey">du-etal-2024-uor</identifier>
<identifier type="doi">10.18653/v1/2024.findings-acl.468</identifier>
<location>
<url>https://aclanthology.org/2024.findings-acl.468</url>
</location>
<part>
<date>2024-08</date>
<extent unit="page">
<start>7865</start>
<end>7877</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T UOR: Universal Backdoor Attacks on Pre-trained Language Models
%A Du, Wei
%A Li, Peixuan
%A Zhao, Haodong
%A Ju, Tianjie
%A Ren, Ge
%A Liu, Gongshen
%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 du-etal-2024-uor
%X Task-agnostic and transferable backdoors implanted in pre-trained language models (PLMs) pose a severe security threat as they can be inherited to any downstream task. However, existing methods rely on manual selection of triggers and backdoor representations, hindering their effectiveness and universality across different PLMs or usage paradigms. In this paper, we propose a new backdoor attack method called UOR, which overcomes these limitations by turning manual selection into automatic optimization. Specifically, we design poisoned supervised contrastive learning, which can automatically learn more uniform and universal backdoor representations. This allows for more even coverage of the output space, thus hitting more labels in downstream tasks after fine-tuning. Furthermore, we utilize gradient search to select appropriate trigger words that can be adapted to different PLMs and vocabularies. Experiments show that UOR achieves better attack performance on various text classification tasks compared to manual methods. Moreover, we test on PLMs with different architectures, usage paradigms, and more challenging tasks, achieving higher scores for universality.
%R 10.18653/v1/2024.findings-acl.468
%U https://aclanthology.org/2024.findings-acl.468
%U https://doi.org/10.18653/v1/2024.findings-acl.468
%P 7865-7877
Markdown (Informal)
[UOR: Universal Backdoor Attacks on Pre-trained Language Models](https://aclanthology.org/2024.findings-acl.468) (Du et al., Findings 2024)
ACL