@inproceedings{lei-etal-2026-moe,
title = "{M}o{E} Adapter for Large Audio Language Models: Sparsity, Disentanglement, and Gradient-Conflict-Free",
author = "Lei, Yishu and
He, Shuwei and
Jing, Hu and
Zhang, Dan and
Luo, Xianlong and
Zhu, Danxiang and
Feng, Shikun and
Liu, Rui and
HE, Jingzhou and
Sun, Yu and
Wu, Hua and
Wang, Haifeng",
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.840/",
pages = "17039--17050",
ISBN = "979-8-89176-395-1",
abstract = "Extending the input modality of Large Language Models (LLMs) to the audio domain is essential for achieving comprehensive multimodal perception. However, it is well-known that acoustic information is intrinsically heterogeneous, entangling attributes such as speech, music, and environmental context. Existing research is limited to a dense, parameter-shared adapter to model these diverse patterns, which induces gradient conflict during optimization, as parameter updates required for distinct attributes contradict each other. To address this limitation, we introduce the MoE-Adapter, a sparse Mixture-of-Experts (MoE) architecture designed to decouple acoustic information. Specifically, it employs a dynamic gating mechanism that routes audio tokens to specialized experts capturing complementary feature subspaces while retaining shared experts for global context, thereby mitigating gradient conflicts and enabling fine-grained feature learning. Comprehensive experiments show that the MoE-Adapter achieves superior performance on both audio semantic and paralinguistic tasks, consistently outperforming dense linear baselines with comparable computational costs. To facilitate future research, our code are publicly available at https://github.com/Alittleegg/Eureka-Audio."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lei-etal-2026-moe">
<titleInfo>
<title>MoE Adapter for Large Audio Language Models: Sparsity, Disentanglement, and Gradient-Conflict-Free</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yishu</namePart>
<namePart type="family">Lei</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shuwei</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hu</namePart>
<namePart type="family">Jing</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dan</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xianlong</namePart>
<namePart type="family">Luo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Danxiang</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shikun</namePart>
<namePart type="family">Feng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rui</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jingzhou</namePart>
<namePart type="family">HE</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yu</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hua</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haifeng</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>Extending the input modality of Large Language Models (LLMs) to the audio domain is essential for achieving comprehensive multimodal perception. However, it is well-known that acoustic information is intrinsically heterogeneous, entangling attributes such as speech, music, and environmental context. Existing research is limited to a dense, parameter-shared adapter to model these diverse patterns, which induces gradient conflict during optimization, as parameter updates required for distinct attributes contradict each other. To address this limitation, we introduce the MoE-Adapter, a sparse Mixture-of-Experts (MoE) architecture designed to decouple acoustic information. Specifically, it employs a dynamic gating mechanism that routes audio tokens to specialized experts capturing complementary feature subspaces while retaining shared experts for global context, thereby mitigating gradient conflicts and enabling fine-grained feature learning. Comprehensive experiments show that the MoE-Adapter achieves superior performance on both audio semantic and paralinguistic tasks, consistently outperforming dense linear baselines with comparable computational costs. To facilitate future research, our code are publicly available at https://github.com/Alittleegg/Eureka-Audio.</abstract>
<identifier type="citekey">lei-etal-2026-moe</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.840/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>17039</start>
<end>17050</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MoE Adapter for Large Audio Language Models: Sparsity, Disentanglement, and Gradient-Conflict-Free
%A Lei, Yishu
%A He, Shuwei
%A Jing, Hu
%A Zhang, Dan
%A Luo, Xianlong
%A Zhu, Danxiang
%A Feng, Shikun
%A Liu, Rui
%A HE, Jingzhou
%A Sun, Yu
%A Wu, Hua
%A Wang, Haifeng
%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 lei-etal-2026-moe
%X Extending the input modality of Large Language Models (LLMs) to the audio domain is essential for achieving comprehensive multimodal perception. However, it is well-known that acoustic information is intrinsically heterogeneous, entangling attributes such as speech, music, and environmental context. Existing research is limited to a dense, parameter-shared adapter to model these diverse patterns, which induces gradient conflict during optimization, as parameter updates required for distinct attributes contradict each other. To address this limitation, we introduce the MoE-Adapter, a sparse Mixture-of-Experts (MoE) architecture designed to decouple acoustic information. Specifically, it employs a dynamic gating mechanism that routes audio tokens to specialized experts capturing complementary feature subspaces while retaining shared experts for global context, thereby mitigating gradient conflicts and enabling fine-grained feature learning. Comprehensive experiments show that the MoE-Adapter achieves superior performance on both audio semantic and paralinguistic tasks, consistently outperforming dense linear baselines with comparable computational costs. To facilitate future research, our code are publicly available at https://github.com/Alittleegg/Eureka-Audio.
%U https://aclanthology.org/2026.findings-acl.840/
%P 17039-17050
Markdown (Informal)
[MoE Adapter for Large Audio Language Models: Sparsity, Disentanglement, and Gradient-Conflict-Free](https://aclanthology.org/2026.findings-acl.840/) (Lei et al., Findings 2026)
ACL
- Yishu Lei, Shuwei He, Hu Jing, Dan Zhang, Xianlong Luo, Danxiang Zhu, Shikun Feng, Rui Liu, Jingzhou HE, Yu Sun, Hua Wu, and Haifeng Wang. 2026. MoE Adapter for Large Audio Language Models: Sparsity, Disentanglement, and Gradient-Conflict-Free. In Findings of the Association for Computational Linguistics: ACL 2026, pages 17039–17050, San Diego, California, United States. Association for Computational Linguistics.