@inproceedings{zhou-etal-2026-beyond-efficiency,
title = "Beyond the Efficiency Paradox: Unlocking Multimodal Subspaces via Cumulative Information Span",
author = "Zhou, Xiuze and
Chu, Xiaowen and
Hu, Xuming",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.2365/",
pages = "53700--53710",
ISBN = "979-8-89176-459-0",
abstract = "In real-world multimodal applications, different modalities often exhibit weakly aligned adaptation requirements, making it challenging for a shared adapter to effectively capture heterogeneous signals. Existing approaches typically use a unified low-rank adapter for concatenated multimodal representations, leading to competition among modality-specific adaptations and diminishing returns as the adapter rank increases. To address this limitation, we propose msLoRA, a modality-specific adaptation framework for efficient multimodal fine-tuning. msLoRA assigns independent low-rank adapters to different modalities, allowing them to capture specialized adaptation directions while maintaining efficient parameterization. Our theoretical analysis introduces the concept of cumulative information span, demonstrating that modality-specific adapters can expand the effective adaptation space and avoid the cross-modal disagreement typical of shared adapters. We evaluate msLoRA using multiple prominent open-source LLMs, including the Llama, Qwen, and Mistral families, across four multimodal tasks, showing consistent improvements in scalability and effectiveness for low-rank adaptation in multimodal LLMs. The source code is available at \url{https://github.com/XiuzeZhou/msLoRA/}."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhou-etal-2026-beyond-efficiency">
<titleInfo>
<title>Beyond the Efficiency Paradox: Unlocking Multimodal Subspaces via Cumulative Information Span</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xiuze</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaowen</namePart>
<namePart type="family">Chu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xuming</namePart>
<namePart type="family">Hu</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</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</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">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>In real-world multimodal applications, different modalities often exhibit weakly aligned adaptation requirements, making it challenging for a shared adapter to effectively capture heterogeneous signals. Existing approaches typically use a unified low-rank adapter for concatenated multimodal representations, leading to competition among modality-specific adaptations and diminishing returns as the adapter rank increases. To address this limitation, we propose msLoRA, a modality-specific adaptation framework for efficient multimodal fine-tuning. msLoRA assigns independent low-rank adapters to different modalities, allowing them to capture specialized adaptation directions while maintaining efficient parameterization. Our theoretical analysis introduces the concept of cumulative information span, demonstrating that modality-specific adapters can expand the effective adaptation space and avoid the cross-modal disagreement typical of shared adapters. We evaluate msLoRA using multiple prominent open-source LLMs, including the Llama, Qwen, and Mistral families, across four multimodal tasks, showing consistent improvements in scalability and effectiveness for low-rank adaptation in multimodal LLMs. The source code is available at https://github.com/XiuzeZhou/msLoRA/.</abstract>
<identifier type="citekey">zhou-etal-2026-beyond-efficiency</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2365/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>53700</start>
<end>53710</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Beyond the Efficiency Paradox: Unlocking Multimodal Subspaces via Cumulative Information Span
%A Zhou, Xiuze
%A Chu, Xiaowen
%A Hu, Xuming
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F zhou-etal-2026-beyond-efficiency
%X In real-world multimodal applications, different modalities often exhibit weakly aligned adaptation requirements, making it challenging for a shared adapter to effectively capture heterogeneous signals. Existing approaches typically use a unified low-rank adapter for concatenated multimodal representations, leading to competition among modality-specific adaptations and diminishing returns as the adapter rank increases. To address this limitation, we propose msLoRA, a modality-specific adaptation framework for efficient multimodal fine-tuning. msLoRA assigns independent low-rank adapters to different modalities, allowing them to capture specialized adaptation directions while maintaining efficient parameterization. Our theoretical analysis introduces the concept of cumulative information span, demonstrating that modality-specific adapters can expand the effective adaptation space and avoid the cross-modal disagreement typical of shared adapters. We evaluate msLoRA using multiple prominent open-source LLMs, including the Llama, Qwen, and Mistral families, across four multimodal tasks, showing consistent improvements in scalability and effectiveness for low-rank adaptation in multimodal LLMs. The source code is available at https://github.com/XiuzeZhou/msLoRA/.
%U https://aclanthology.org/2026.emnlp-main.2365/
%P 53700-53710
Markdown (Informal)
[Beyond the Efficiency Paradox: Unlocking Multimodal Subspaces via Cumulative Information Span](https://aclanthology.org/2026.emnlp-main.2365/) (Zhou et al., EMNLP 2026)
ACL