<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhou-etal-2022-sharpness">
    <titleInfo>
        <title>Sharpness-Aware Minimization with Dynamic Reweighting</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Wenxuan</namePart>
        <namePart type="family">Zhou</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Fangyu</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Huan</namePart>
        <namePart type="family">Zhang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Muhao</namePart>
        <namePart type="family">Chen</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2022-12</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: EMNLP 2022</title>
        </titleInfo>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Abu Dhabi, United Arab Emirates</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>Deep neural networks are often overparameterized and may not easily achieve model generalization. Adversarial training has shown effectiveness in improving generalization by regularizing the change of loss on top of adversarially chosen perturbations. The recently proposed sharpness-aware minimization (SAM) algorithm conducts adversarial weight perturbation, encouraging the model to converge to a flat minima. SAM finds a common adversarial weight perturbation per-batch. Although per-instance adversarial weight perturbations are stronger adversaries and can potentially lead to better generalization performance, their computational cost is very high and thus it is impossible to use per-instance perturbations efficiently in SAM. In this paper, we tackle this efficiency bottleneck and propose sharpness-aware minimization with dynamic reweighting (delta-SAM). Our theoretical analysis motivates that it is possible to approach the stronger, per-instance adversarial weight perturbations using reweighted per-batch weight perturbations. delta-SAM dynamically reweights perturbation within each batch according to the theoretically principled weighting factors, serving as a good approximation to per-instance perturbation. Experiments on various natural language understanding tasks demonstrate the effectiveness of delta-SAM.</abstract>
    <identifier type="citekey">zhou-etal-2022-sharpness</identifier>
    <location>
        <url>https://aclanthology.org/2022.findings-emnlp.417</url>
    </location>
    <part>
        <date>2022-12</date>
        <extent unit="page">
            <start>5686</start>
            <end>5699</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Sharpness-Aware Minimization with Dynamic Reweighting
%A Zhou, Wenxuan
%A Liu, Fangyu
%A Zhang, Huan
%A Chen, Muhao
%S Findings of the Association for Computational Linguistics: EMNLP 2022
%D 2022
%8 December
%I Association for Computational Linguistics
%C Abu Dhabi, United Arab Emirates
%F zhou-etal-2022-sharpness
%X Deep neural networks are often overparameterized and may not easily achieve model generalization. Adversarial training has shown effectiveness in improving generalization by regularizing the change of loss on top of adversarially chosen perturbations. The recently proposed sharpness-aware minimization (SAM) algorithm conducts adversarial weight perturbation, encouraging the model to converge to a flat minima. SAM finds a common adversarial weight perturbation per-batch. Although per-instance adversarial weight perturbations are stronger adversaries and can potentially lead to better generalization performance, their computational cost is very high and thus it is impossible to use per-instance perturbations efficiently in SAM. In this paper, we tackle this efficiency bottleneck and propose sharpness-aware minimization with dynamic reweighting (delta-SAM). Our theoretical analysis motivates that it is possible to approach the stronger, per-instance adversarial weight perturbations using reweighted per-batch weight perturbations. delta-SAM dynamically reweights perturbation within each batch according to the theoretically principled weighting factors, serving as a good approximation to per-instance perturbation. Experiments on various natural language understanding tasks demonstrate the effectiveness of delta-SAM.
%U https://aclanthology.org/2022.findings-emnlp.417
%P 5686-5699

Markdown (Informal)

[Sharpness-Aware Minimization with Dynamic Reweighting](https://aclanthology.org/2022.findings-emnlp.417) (Zhou et al., Findings 2022)

• Sharpness-Aware Minimization with Dynamic Reweighting (Zhou et al., Findings 2022)

ACL

• Wenxuan Zhou, Fangyu Liu, Huan Zhang, and Muhao Chen. 2022. Sharpness-Aware Minimization with Dynamic Reweighting. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 5686–5699, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.