@inproceedings{park-etal-2026-dplora,
title = "{DPL}o{RA}: A Dual-Pruning Framework based on {ILP} Optimization and Progressive Pruning for Parameter-Efficient {L}o{RA} Fine-Tuning",
author = "Park, Changjun and
Yoon, Sejong and
Kim, Jaekwang",
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.1693/",
doi = "10.18653/v1/2026.findings-acl.1693",
pages = "33892--33908",
ISBN = "979-8-89176-395-1",
abstract = "We propose DPLoRA (Dual-Pruning Low-Rank Adaptation), a framework that optimizes rank allocation via two stages: (1) an initial pruning stage (OPLoRA; Optimal Pruning LoRA) that uses Integer Linear Programming (ILP) to determine optimal layer-wise ranks without manual tuning; and (2) a progressive pruning stage that further reduces ranks adaptively during training using importance scores smoothed by Exponential Moving Average (EMA). Experiments demonstrate that OPLoRA consistently outperforms existing PEFT baselines on GLUE and instruction-following tasks, while the full DPLoRA framework establishes a new state-of-the-art among compared PEFT baselines on GLUE at high-performance settings ($p=0.4$). At efficiency-focused settings ($p=0.8$), our method reduces trainable parameters by over 80{\%} and training time by 46{\%} compared to standard LoRA, offering a highly efficient solution for deploying large-scale models in resource-constrained environments."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="park-etal-2026-dplora">
<titleInfo>
<title>DPLoRA: A Dual-Pruning Framework based on ILP Optimization and Progressive Pruning for Parameter-Efficient LoRA Fine-Tuning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Changjun</namePart>
<namePart type="family">Park</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sejong</namePart>
<namePart type="family">Yoon</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jaekwang</namePart>
<namePart type="family">Kim</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>We propose DPLoRA (Dual-Pruning Low-Rank Adaptation), a framework that optimizes rank allocation via two stages: (1) an initial pruning stage (OPLoRA; Optimal Pruning LoRA) that uses Integer Linear Programming (ILP) to determine optimal layer-wise ranks without manual tuning; and (2) a progressive pruning stage that further reduces ranks adaptively during training using importance scores smoothed by Exponential Moving Average (EMA). Experiments demonstrate that OPLoRA consistently outperforms existing PEFT baselines on GLUE and instruction-following tasks, while the full DPLoRA framework establishes a new state-of-the-art among compared PEFT baselines on GLUE at high-performance settings (p=0.4). At efficiency-focused settings (p=0.8), our method reduces trainable parameters by over 80% and training time by 46% compared to standard LoRA, offering a highly efficient solution for deploying large-scale models in resource-constrained environments.</abstract>
<identifier type="citekey">park-etal-2026-dplora</identifier>
<identifier type="doi">10.18653/v1/2026.findings-acl.1693</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.1693/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>33892</start>
<end>33908</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T DPLoRA: A Dual-Pruning Framework based on ILP Optimization and Progressive Pruning for Parameter-Efficient LoRA Fine-Tuning
%A Park, Changjun
%A Yoon, Sejong
%A Kim, Jaekwang
%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 park-etal-2026-dplora
%X We propose DPLoRA (Dual-Pruning Low-Rank Adaptation), a framework that optimizes rank allocation via two stages: (1) an initial pruning stage (OPLoRA; Optimal Pruning LoRA) that uses Integer Linear Programming (ILP) to determine optimal layer-wise ranks without manual tuning; and (2) a progressive pruning stage that further reduces ranks adaptively during training using importance scores smoothed by Exponential Moving Average (EMA). Experiments demonstrate that OPLoRA consistently outperforms existing PEFT baselines on GLUE and instruction-following tasks, while the full DPLoRA framework establishes a new state-of-the-art among compared PEFT baselines on GLUE at high-performance settings (p=0.4). At efficiency-focused settings (p=0.8), our method reduces trainable parameters by over 80% and training time by 46% compared to standard LoRA, offering a highly efficient solution for deploying large-scale models in resource-constrained environments.
%R 10.18653/v1/2026.findings-acl.1693
%U https://aclanthology.org/2026.findings-acl.1693/
%U https://doi.org/10.18653/v1/2026.findings-acl.1693
%P 33892-33908
Markdown (Informal)
[DPLoRA: A Dual-Pruning Framework based on ILP Optimization and Progressive Pruning for Parameter-Efficient LoRA Fine-Tuning](https://aclanthology.org/2026.findings-acl.1693/) (Park et al., Findings 2026)
ACL