<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="xu-etal-2026-fedanchor">
    <titleInfo>
        <title>FedAnchor: Differentially Private Federated LoRA via Weight Anchoring</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Jie</namePart>
        <namePart type="family">Xu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Haaris</namePart>
        <namePart type="family">Mehmood</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Rogier</namePart>
        <namePart type="family">van Dalen</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Karthikeyan</namePart>
        <namePart type="family">Saravanan</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Mete</namePart>
        <namePart type="family">Ozay</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Pepa</namePart>
            <namePart type="family">Atanasova</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Gholamreza</namePart>
            <namePart type="family">Haffari</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Minghan</namePart>
            <namePart type="family">Wang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-460-6</identifier>
    </relatedItem>
    <abstract>Federated learning (FL) with differential privacy (DP) allows several organizations to learn a shared model from user data that only one organization is allowed to see. However, guaranteeing DP requires adding noise to gradients that scales with update dimensionality. While low-rank adaptation (LoRA) is a popular method for parameter-efficient fine-tuning, it is inherently sensitive to DP noise due to the multiplication operator. This noise amplification leads to significant utility degradation, hindering practical deployment in enterprise settings. To overcome this, we propose FedAnchor. Every round, FedAnchor captures compressed information using one of the LoRA matrices A and B, and then decompresses this to the complete adapted weight matrix via singular value decomposition (SVD) at the server. This relieves the model from DP noise exacerbation while maintaining full learning capacity. Experiments across multiple benchmarks demonstrate that FedAnchor matches the strongest baseline in non-private settings and outperforms it by 4.38% under stringent DP constraints (ε = 3). At scale, FedAnchor maintains communication efficiency, computation robustness, and utility under strict privacy budgets, meeting the demands for enterprise-level cross-silo deployments.</abstract>
    <identifier type="citekey">xu-etal-2026-fedanchor</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-industry.211/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>3544</start>
            <end>3554</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T FedAnchor: Differentially Private Federated LoRA via Weight Anchoring
%A Xu, Jie
%A Mehmood, Haaris
%A van Dalen, Rogier
%A Saravanan, Karthikeyan
%A Ozay, Mete
%Y Atanasova, Pepa
%Y Haffari, Gholamreza
%Y Wang, Minghan
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-460-6
%F xu-etal-2026-fedanchor
%X Federated learning (FL) with differential privacy (DP) allows several organizations to learn a shared model from user data that only one organization is allowed to see. However, guaranteeing DP requires adding noise to gradients that scales with update dimensionality. While low-rank adaptation (LoRA) is a popular method for parameter-efficient fine-tuning, it is inherently sensitive to DP noise due to the multiplication operator. This noise amplification leads to significant utility degradation, hindering practical deployment in enterprise settings. To overcome this, we propose FedAnchor. Every round, FedAnchor captures compressed information using one of the LoRA matrices A and B, and then decompresses this to the complete adapted weight matrix via singular value decomposition (SVD) at the server. This relieves the model from DP noise exacerbation while maintaining full learning capacity. Experiments across multiple benchmarks demonstrate that FedAnchor matches the strongest baseline in non-private settings and outperforms it by 4.38% under stringent DP constraints (ε = 3). At scale, FedAnchor maintains communication efficiency, computation robustness, and utility under strict privacy budgets, meeting the demands for enterprise-level cross-silo deployments.
%U https://aclanthology.org/2026.emnlp-industry.211/
%P 3544-3554

Markdown (Informal)

[FedAnchor: Differentially Private Federated LoRA via Weight Anchoring](https://aclanthology.org/2026.emnlp-industry.211/) (Xu et al., EMNLP 2026)

• FedAnchor: Differentially Private Federated LoRA via Weight Anchoring (Xu et al., EMNLP 2026)

ACL

• Jie Xu, Haaris Mehmood, Rogier van Dalen, Karthikeyan Saravanan, and Mete Ozay. 2026. FedAnchor: Differentially Private Federated LoRA via Weight Anchoring. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track, pages 3544–3554, Budapest, Hungary. Association for Computational Linguistics.