@inproceedings{yoon-choi-2026-compel,
title = "{COMPEL}: Compensated Mixture-of-Experts Pruning with Expert-Layer distribution",
author = "Yoon, Seohee and
Choi, Yong Suk",
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.1521/",
pages = "30408--30423",
ISBN = "979-8-89176-395-1",
abstract = "Mixture-of-Experts (MoE) architectures have emerged as an effective approach for scaling Large Language Models (LLMs) by activating only a subset of experts during inference. Despite their computational efficiency, MoE models incur a substantial memory bottleneck from maintaining all expert parameters during inference. To address this challenge, numerous MoE pruning methods have been proposed. However, most existing methods adopt uniform pruning across layers, which fails to capture layer-wise variations in expert importance and redundancy. In this paper, we propose COmpensated MoE Pruning with Expert-Layer distribution (COMPEL). COMPEL performs layer-adaptive expert pruning by estimating expert importance using Fisher information and deriving layer importance from layer-wise outlier distributions, enabling pruning decisions that capture layer-wise heterogeneity. Furthermore, to mitigate performance degradation resulting from expert pruning, we propose a Fisher information guided expert weight compensation method. Experimental results on the Qwen1.5-MoE-A2.7B achieve near lossless performance at 25{\%} expert pruning and maintains performance within a 4{\%} margin even at 50{\%} pruning. Moreover, COMPEL consistently outperforms existing pruning methods while substantially reducing inference latency and peak GPU memory usage."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yoon-choi-2026-compel">
<titleInfo>
<title>COMPEL: Compensated Mixture-of-Experts Pruning with Expert-Layer distribution</title>
</titleInfo>
<name type="personal">
<namePart type="given">Seohee</namePart>
<namePart type="family">Yoon</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yong</namePart>
<namePart type="given">Suk</namePart>
<namePart type="family">Choi</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>Mixture-of-Experts (MoE) architectures have emerged as an effective approach for scaling Large Language Models (LLMs) by activating only a subset of experts during inference. Despite their computational efficiency, MoE models incur a substantial memory bottleneck from maintaining all expert parameters during inference. To address this challenge, numerous MoE pruning methods have been proposed. However, most existing methods adopt uniform pruning across layers, which fails to capture layer-wise variations in expert importance and redundancy. In this paper, we propose COmpensated MoE Pruning with Expert-Layer distribution (COMPEL). COMPEL performs layer-adaptive expert pruning by estimating expert importance using Fisher information and deriving layer importance from layer-wise outlier distributions, enabling pruning decisions that capture layer-wise heterogeneity. Furthermore, to mitigate performance degradation resulting from expert pruning, we propose a Fisher information guided expert weight compensation method. Experimental results on the Qwen1.5-MoE-A2.7B achieve near lossless performance at 25% expert pruning and maintains performance within a 4% margin even at 50% pruning. Moreover, COMPEL consistently outperforms existing pruning methods while substantially reducing inference latency and peak GPU memory usage.</abstract>
<identifier type="citekey">yoon-choi-2026-compel</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.1521/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>30408</start>
<end>30423</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T COMPEL: Compensated Mixture-of-Experts Pruning with Expert-Layer distribution
%A Yoon, Seohee
%A Choi, Yong Suk
%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 yoon-choi-2026-compel
%X Mixture-of-Experts (MoE) architectures have emerged as an effective approach for scaling Large Language Models (LLMs) by activating only a subset of experts during inference. Despite their computational efficiency, MoE models incur a substantial memory bottleneck from maintaining all expert parameters during inference. To address this challenge, numerous MoE pruning methods have been proposed. However, most existing methods adopt uniform pruning across layers, which fails to capture layer-wise variations in expert importance and redundancy. In this paper, we propose COmpensated MoE Pruning with Expert-Layer distribution (COMPEL). COMPEL performs layer-adaptive expert pruning by estimating expert importance using Fisher information and deriving layer importance from layer-wise outlier distributions, enabling pruning decisions that capture layer-wise heterogeneity. Furthermore, to mitigate performance degradation resulting from expert pruning, we propose a Fisher information guided expert weight compensation method. Experimental results on the Qwen1.5-MoE-A2.7B achieve near lossless performance at 25% expert pruning and maintains performance within a 4% margin even at 50% pruning. Moreover, COMPEL consistently outperforms existing pruning methods while substantially reducing inference latency and peak GPU memory usage.
%U https://aclanthology.org/2026.findings-acl.1521/
%P 30408-30423
Markdown (Informal)
[COMPEL: Compensated Mixture-of-Experts Pruning with Expert-Layer distribution](https://aclanthology.org/2026.findings-acl.1521/) (Yoon & Choi, Findings 2026)
ACL