@inproceedings{wang-etal-2026-federated,
title = "Federated {L}o{RA} Fine-Tuning with Pipelined Error-Mitigated Aggregation and Matrix-Wise Freezing",
author = "Wang, Haoran and
Wang, Xiong and
Li, Yuqing and
Chen, Jing and
Zhang, Junyi and
Yan, Nan and
He, Kun and
Wang, Wei",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.284/",
pages = "5749--5762",
ISBN = "979-8-89176-395-1",
abstract = "Federated low-rank adaptation (LoRA) enables multiple clients to collaboratively fine-tune large language models (LLMs) without disclosing their raw data. However, existing works often experience performance degradation due to biased model aggregation and are hindered by significant communication and computation burden, both limiting training efficiency. In this paper, we propose iFLoRA, an improved Federated LoRA fine-tuning system for LLMs featuring pipelined error-mitigated model aggregation and adaptive matrix-wise parameter freezing. Specifically, iFLoRA mitigates aggregation error by first reconstructing local update matrices from clients' low-rank matrices. These are then aggregated into a global update, which is decomposed via singular value decomposition (SVD) to form low-rank matrices for the next round. To mitigate the overhead from SVD, iFLoRA employs a pipeline to overlap global aggregation, local computation, and communication. Additionally, iFLoRA implements an adaptive matrix-wise freezing scheme that assesses their stability and selectively freezes them for adaptively adjusted periods, alleviating client training overheads without compromising model performance. Extensive experiments on real-world datasets show that iFLoRA can improve time-to-target by 2.17-8.48{\texttimes} than state-of-the-art methods. Our code is available at: https://github.com/whr819987540/iflora."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2026-federated">
<titleInfo>
<title>Federated LoRA Fine-Tuning with Pipelined Error-Mitigated Aggregation and Matrix-Wise Freezing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Haoran</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiong</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuqing</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jing</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junyi</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nan</namePart>
<namePart type="family">Yan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kun</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wei</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>Federated low-rank adaptation (LoRA) enables multiple clients to collaboratively fine-tune large language models (LLMs) without disclosing their raw data. However, existing works often experience performance degradation due to biased model aggregation and are hindered by significant communication and computation burden, both limiting training efficiency. In this paper, we propose iFLoRA, an improved Federated LoRA fine-tuning system for LLMs featuring pipelined error-mitigated model aggregation and adaptive matrix-wise parameter freezing. Specifically, iFLoRA mitigates aggregation error by first reconstructing local update matrices from clients’ low-rank matrices. These are then aggregated into a global update, which is decomposed via singular value decomposition (SVD) to form low-rank matrices for the next round. To mitigate the overhead from SVD, iFLoRA employs a pipeline to overlap global aggregation, local computation, and communication. Additionally, iFLoRA implements an adaptive matrix-wise freezing scheme that assesses their stability and selectively freezes them for adaptively adjusted periods, alleviating client training overheads without compromising model performance. Extensive experiments on real-world datasets show that iFLoRA can improve time-to-target by 2.17-8.48× than state-of-the-art methods. Our code is available at: https://github.com/whr819987540/iflora.</abstract>
<identifier type="citekey">wang-etal-2026-federated</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.284/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>5749</start>
<end>5762</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Federated LoRA Fine-Tuning with Pipelined Error-Mitigated Aggregation and Matrix-Wise Freezing
%A Wang, Haoran
%A Wang, Xiong
%A Li, Yuqing
%A Chen, Jing
%A Zhang, Junyi
%A Yan, Nan
%A He, Kun
%A Wang, Wei
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F wang-etal-2026-federated
%X Federated low-rank adaptation (LoRA) enables multiple clients to collaboratively fine-tune large language models (LLMs) without disclosing their raw data. However, existing works often experience performance degradation due to biased model aggregation and are hindered by significant communication and computation burden, both limiting training efficiency. In this paper, we propose iFLoRA, an improved Federated LoRA fine-tuning system for LLMs featuring pipelined error-mitigated model aggregation and adaptive matrix-wise parameter freezing. Specifically, iFLoRA mitigates aggregation error by first reconstructing local update matrices from clients’ low-rank matrices. These are then aggregated into a global update, which is decomposed via singular value decomposition (SVD) to form low-rank matrices for the next round. To mitigate the overhead from SVD, iFLoRA employs a pipeline to overlap global aggregation, local computation, and communication. Additionally, iFLoRA implements an adaptive matrix-wise freezing scheme that assesses their stability and selectively freezes them for adaptively adjusted periods, alleviating client training overheads without compromising model performance. Extensive experiments on real-world datasets show that iFLoRA can improve time-to-target by 2.17-8.48× than state-of-the-art methods. Our code is available at: https://github.com/whr819987540/iflora.
%U https://aclanthology.org/2026.findings-acl.284/
%P 5749-5762
Markdown (Informal)
[Federated LoRA Fine-Tuning with Pipelined Error-Mitigated Aggregation and Matrix-Wise Freezing](https://aclanthology.org/2026.findings-acl.284/) (Wang et al., Findings 2026)
ACL
- Haoran Wang, Xiong Wang, Yuqing Li, Jing Chen, Junyi Zhang, Nan Yan, Kun He, and Wei Wang. 2026. Federated LoRA Fine-Tuning with Pipelined Error-Mitigated Aggregation and Matrix-Wise Freezing. In Findings of the Association for Computational Linguistics: ACL 2026, pages 5749–5762, San Diego, California, United States. Association for Computational Linguistics.