<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wu-etal-2023-depn">
    <titleInfo>
        <title>DEPN: Detecting and Editing Privacy Neurons in Pretrained Language Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Xinwei</namePart>
        <namePart type="family">Wu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Junzhuo</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Minghui</namePart>
        <namePart type="family">Xu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Weilong</namePart>
        <namePart type="family">Dong</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Shuangzhi</namePart>
        <namePart type="family">Wu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Chao</namePart>
        <namePart type="family">Bian</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Deyi</namePart>
        <namePart type="family">Xiong</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>Pretrained language models have learned a vast amount of human knowledge from large-scale corpora, but their powerful memorization capability also brings the risk of data leakage. Some risks may only be discovered after the model training is completed, such as the model memorizing a specific phone number and frequently outputting it. In such cases, model developers need to eliminate specific data influences from the model to mitigate legal and ethical penalties. To effectively mitigate these risks, people often have to spend a significant amount of time and computational costs to retrain new models instead of finding ways to cure the ‘sick’ models. Therefore, we propose a method to locate and erase risky neurons in order to eliminate the impact of privacy data in the model. We use a new method based on integrated gradients to locate neurons associated with privacy texts, and then erase these neurons by setting their activation values to zero.Furthermore, we propose a risky neuron aggregation method to eliminate the influence of privacy data in the model in batches. Experimental results show that our method can effectively and quickly eliminate the impact of privacy data without affecting the model’s performance. Additionally, we demonstrate the relationship between model memorization and neurons through experiments, further illustrating the robustness of our method.</abstract>
    <identifier type="citekey">wu-etal-2023-depn</identifier>
    <identifier type="doi">10.18653/v1/2023.emnlp-main.174</identifier>
    <location>
        <url>https://aclanthology.org/2023.emnlp-main.174</url>
    </location>
    <part>
        <date>2023-12</date>
        <extent unit="page">
            <start>2875</start>
            <end>2886</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T DEPN: Detecting and Editing Privacy Neurons in Pretrained Language Models
%A Wu, Xinwei
%A Li, Junzhuo
%A Xu, Minghui
%A Dong, Weilong
%A Wu, Shuangzhi
%A Bian, Chao
%A Xiong, Deyi
%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 wu-etal-2023-depn
%X Pretrained language models have learned a vast amount of human knowledge from large-scale corpora, but their powerful memorization capability also brings the risk of data leakage. Some risks may only be discovered after the model training is completed, such as the model memorizing a specific phone number and frequently outputting it. In such cases, model developers need to eliminate specific data influences from the model to mitigate legal and ethical penalties. To effectively mitigate these risks, people often have to spend a significant amount of time and computational costs to retrain new models instead of finding ways to cure the ‘sick’ models. Therefore, we propose a method to locate and erase risky neurons in order to eliminate the impact of privacy data in the model. We use a new method based on integrated gradients to locate neurons associated with privacy texts, and then erase these neurons by setting their activation values to zero.Furthermore, we propose a risky neuron aggregation method to eliminate the influence of privacy data in the model in batches. Experimental results show that our method can effectively and quickly eliminate the impact of privacy data without affecting the model’s performance. Additionally, we demonstrate the relationship between model memorization and neurons through experiments, further illustrating the robustness of our method.
%R 10.18653/v1/2023.emnlp-main.174
%U https://aclanthology.org/2023.emnlp-main.174
%U https://doi.org/10.18653/v1/2023.emnlp-main.174
%P 2875-2886

Markdown (Informal)

[DEPN: Detecting and Editing Privacy Neurons in Pretrained Language Models](https://aclanthology.org/2023.emnlp-main.174) (Wu et al., EMNLP 2023)

• DEPN: Detecting and Editing Privacy Neurons in Pretrained Language Models (Wu et al., EMNLP 2023)

ACL

• Xinwei Wu, Junzhuo Li, Minghui Xu, Weilong Dong, Shuangzhi Wu, Chao Bian, and Deyi Xiong. 2023. DEPN: Detecting and Editing Privacy Neurons in Pretrained Language Models. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 2875–2886, Singapore. Association for Computational Linguistics.