@inproceedings{zhang-etal-2025-diversifying,
title = "Diversifying the Expert Knowledge for Task-Agnostic Pruning in Sparse Mixture-of-Experts",
author = "Zhang, Zeliang and
Liu, Xiaodong and
Cheng, Hao and
Xu, Chenliang and
Gao, Jianfeng",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-acl.4/",
doi = "10.18653/v1/2025.findings-acl.4",
pages = "86--102",
ISBN = "979-8-89176-256-5",
abstract = "In this work, we address the memory overhead of deploying Mixture-of-Experts (MoE) architectures in Large Language Models (LLMs). While MoE layers improve LLM performance without increasing inference costs, the ever-growing number of experts inflates memory requirements, hindering practical deployment. Our empirical study reveals that some experts encode redundant knowledge during pre-training. We thus propose a method of grouping and pruning similar experts to improve the model{'}s parameter efficiency. We validate the effectiveness of our method by pruning three state-of-the-art MoE architectures, including Mixtral, Deepseek-MoE, and Qwen. The evaluation shows that our method outperforms other model pruning methods on a range of natural language tasks."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2025-diversifying">
<titleInfo>
<title>Diversifying the Expert Knowledge for Task-Agnostic Pruning in Sparse Mixture-of-Experts</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zeliang</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaodong</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hao</namePart>
<namePart type="family">Cheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chenliang</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jianfeng</namePart>
<namePart type="family">Gao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-256-5</identifier>
</relatedItem>
<abstract>In this work, we address the memory overhead of deploying Mixture-of-Experts (MoE) architectures in Large Language Models (LLMs). While MoE layers improve LLM performance without increasing inference costs, the ever-growing number of experts inflates memory requirements, hindering practical deployment. Our empirical study reveals that some experts encode redundant knowledge during pre-training. We thus propose a method of grouping and pruning similar experts to improve the model’s parameter efficiency. We validate the effectiveness of our method by pruning three state-of-the-art MoE architectures, including Mixtral, Deepseek-MoE, and Qwen. The evaluation shows that our method outperforms other model pruning methods on a range of natural language tasks.</abstract>
<identifier type="citekey">zhang-etal-2025-diversifying</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.4</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.4/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>86</start>
<end>102</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Diversifying the Expert Knowledge for Task-Agnostic Pruning in Sparse Mixture-of-Experts
%A Zhang, Zeliang
%A Liu, Xiaodong
%A Cheng, Hao
%A Xu, Chenliang
%A Gao, Jianfeng
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F zhang-etal-2025-diversifying
%X In this work, we address the memory overhead of deploying Mixture-of-Experts (MoE) architectures in Large Language Models (LLMs). While MoE layers improve LLM performance without increasing inference costs, the ever-growing number of experts inflates memory requirements, hindering practical deployment. Our empirical study reveals that some experts encode redundant knowledge during pre-training. We thus propose a method of grouping and pruning similar experts to improve the model’s parameter efficiency. We validate the effectiveness of our method by pruning three state-of-the-art MoE architectures, including Mixtral, Deepseek-MoE, and Qwen. The evaluation shows that our method outperforms other model pruning methods on a range of natural language tasks.
%R 10.18653/v1/2025.findings-acl.4
%U https://aclanthology.org/2025.findings-acl.4/
%U https://doi.org/10.18653/v1/2025.findings-acl.4
%P 86-102
Markdown (Informal)
[Diversifying the Expert Knowledge for Task-Agnostic Pruning in Sparse Mixture-of-Experts](https://aclanthology.org/2025.findings-acl.4/) (Zhang et al., Findings 2025)
ACL