@inproceedings{zhou-etal-2026-balancing,
title = "Balancing Fidelity and Plasticity: Aligning Mixed-Precision Fine-Tuning with Linguistic Hierarchies",
author = "Zhou, Changhai and
Zhang, Shiyang and
Zhou, Yuhua and
Gao, Jun and
Qiao, Qian and
Weng, Shichao and
Zhang, Weizhong and
Jin, Cheng",
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.779/",
pages = "15885--15896",
ISBN = "979-8-89176-395-1",
abstract = "Deploying and fine-tuning Large Language Models (LLMs) on resource-constrained edge devices requires navigating a strict trade-off between memory footprint and task performance. Existing quantization-aware fine-tuning methods typically decouple weight precision and adapter capacity, overlooking that a layer{'}s ability to adapt is constrained by the information preserved in its frozen weights. Layers that are highly sensitive to quantization{---}whether due to representational specialization or accumulated error propagation{---}can become bottlenecks that adapter rank alone cannot recover. To address this issue, we introduce \textbf{QR-Adaptor}, a unified framework that jointly optimizes per-layer quantization bit-width and LoRA rank. We formulate resource allocation as a multi-objective discrete search guided by empirical layer-wise sensitivity, and implement it with a three-stage pipeline comprising KL-based sensitivity profiling, evolutionary exploration, and Bayesian refinement. Extensive experiments across LLaMA and Qwen models, including modern instruction tuning on OpenOrca and comparisons with strong PEFT baselines such as QDoRA, show that QR-Adaptor establishes a strong Pareto frontier: under a strict 4-bit memory budget, it matches or approaches 16-bit baselines while using substantially less memory."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhou-etal-2026-balancing">
<titleInfo>
<title>Balancing Fidelity and Plasticity: Aligning Mixed-Precision Fine-Tuning with Linguistic Hierarchies</title>
</titleInfo>
<name type="personal">
<namePart type="given">Changhai</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shiyang</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuhua</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Gao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qian</namePart>
<namePart type="family">Qiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shichao</namePart>
<namePart type="family">Weng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weizhong</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Cheng</namePart>
<namePart type="family">Jin</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>Deploying and fine-tuning Large Language Models (LLMs) on resource-constrained edge devices requires navigating a strict trade-off between memory footprint and task performance. Existing quantization-aware fine-tuning methods typically decouple weight precision and adapter capacity, overlooking that a layer’s ability to adapt is constrained by the information preserved in its frozen weights. Layers that are highly sensitive to quantization—whether due to representational specialization or accumulated error propagation—can become bottlenecks that adapter rank alone cannot recover. To address this issue, we introduce QR-Adaptor, a unified framework that jointly optimizes per-layer quantization bit-width and LoRA rank. We formulate resource allocation as a multi-objective discrete search guided by empirical layer-wise sensitivity, and implement it with a three-stage pipeline comprising KL-based sensitivity profiling, evolutionary exploration, and Bayesian refinement. Extensive experiments across LLaMA and Qwen models, including modern instruction tuning on OpenOrca and comparisons with strong PEFT baselines such as QDoRA, show that QR-Adaptor establishes a strong Pareto frontier: under a strict 4-bit memory budget, it matches or approaches 16-bit baselines while using substantially less memory.</abstract>
<identifier type="citekey">zhou-etal-2026-balancing</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.779/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>15885</start>
<end>15896</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Balancing Fidelity and Plasticity: Aligning Mixed-Precision Fine-Tuning with Linguistic Hierarchies
%A Zhou, Changhai
%A Zhang, Shiyang
%A Zhou, Yuhua
%A Gao, Jun
%A Qiao, Qian
%A Weng, Shichao
%A Zhang, Weizhong
%A Jin, Cheng
%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 zhou-etal-2026-balancing
%X Deploying and fine-tuning Large Language Models (LLMs) on resource-constrained edge devices requires navigating a strict trade-off between memory footprint and task performance. Existing quantization-aware fine-tuning methods typically decouple weight precision and adapter capacity, overlooking that a layer’s ability to adapt is constrained by the information preserved in its frozen weights. Layers that are highly sensitive to quantization—whether due to representational specialization or accumulated error propagation—can become bottlenecks that adapter rank alone cannot recover. To address this issue, we introduce QR-Adaptor, a unified framework that jointly optimizes per-layer quantization bit-width and LoRA rank. We formulate resource allocation as a multi-objective discrete search guided by empirical layer-wise sensitivity, and implement it with a three-stage pipeline comprising KL-based sensitivity profiling, evolutionary exploration, and Bayesian refinement. Extensive experiments across LLaMA and Qwen models, including modern instruction tuning on OpenOrca and comparisons with strong PEFT baselines such as QDoRA, show that QR-Adaptor establishes a strong Pareto frontier: under a strict 4-bit memory budget, it matches or approaches 16-bit baselines while using substantially less memory.
%U https://aclanthology.org/2026.findings-acl.779/
%P 15885-15896
Markdown (Informal)
[Balancing Fidelity and Plasticity: Aligning Mixed-Precision Fine-Tuning with Linguistic Hierarchies](https://aclanthology.org/2026.findings-acl.779/) (Zhou et al., Findings 2026)
ACL
- Changhai Zhou, Shiyang Zhang, Yuhua Zhou, Jun Gao, Qian Qiao, Shichao Weng, Weizhong Zhang, and Cheng Jin. 2026. Balancing Fidelity and Plasticity: Aligning Mixed-Precision Fine-Tuning with Linguistic Hierarchies. In Findings of the Association for Computational Linguistics: ACL 2026, pages 15885–15896, San Diego, California, United States. Association for Computational Linguistics.