@inproceedings{luo-etal-2026-lightmoe,
title = "{L}ight{M}o{E}: Task-Aware Expert Availability Management for Memory-Efficient {M}o{E}-{LLM} Inference",
author = "Luo, Puhan and
Yao, Yunhao and
Wang, Junyang and
Zhang, Junyang and
Li, Xiangyang",
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.1062/",
pages = "21124--21134",
ISBN = "979-8-89176-395-1",
abstract = "Mixture-of-Experts (MoE) models offer a promising path for scaling model capacity, yet their massive memory footprint poses significant challenges for deployment on resource-constrained edge devices. Existing solutions, such as static pruning or dynamic offloading, often struggle to balance model accuracy with inference latency due to irreversible information loss or prohibitive I/O overhead. In this paper, we propose LightMoE, a novel framework for memory-efficient MoE inference that exploits the inherent functional redundancy and temporal locality of expert activation. LightMoE employs a frequency-aware expert initialization strategy to retain a compact core of resident experts and introduces a similarity-based redirection mechanism to compensate for missing experts without incurring I/O costs. Furthermore, it incorporates a lightweight runtime manager that performs coarse-grained, task-level expert replacement to adapt to shifting data distributions. Empirical evaluations on representative edge platforms demonstrate that LightMoE achieves a superior accuracy-efficiency trade-off, improving average accuracy by 4.3{\%} over static pruning and 2.4{\%} over dynamic swapping methods, while maintaining inference latency comparable to strictly pruned models."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="luo-etal-2026-lightmoe">
<titleInfo>
<title>LightMoE: Task-Aware Expert Availability Management for Memory-Efficient MoE-LLM Inference</title>
</titleInfo>
<name type="personal">
<namePart type="given">Puhan</namePart>
<namePart type="family">Luo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yunhao</namePart>
<namePart type="family">Yao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junyang</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junyang</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiangyang</namePart>
<namePart type="family">Li</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) models offer a promising path for scaling model capacity, yet their massive memory footprint poses significant challenges for deployment on resource-constrained edge devices. Existing solutions, such as static pruning or dynamic offloading, often struggle to balance model accuracy with inference latency due to irreversible information loss or prohibitive I/O overhead. In this paper, we propose LightMoE, a novel framework for memory-efficient MoE inference that exploits the inherent functional redundancy and temporal locality of expert activation. LightMoE employs a frequency-aware expert initialization strategy to retain a compact core of resident experts and introduces a similarity-based redirection mechanism to compensate for missing experts without incurring I/O costs. Furthermore, it incorporates a lightweight runtime manager that performs coarse-grained, task-level expert replacement to adapt to shifting data distributions. Empirical evaluations on representative edge platforms demonstrate that LightMoE achieves a superior accuracy-efficiency trade-off, improving average accuracy by 4.3% over static pruning and 2.4% over dynamic swapping methods, while maintaining inference latency comparable to strictly pruned models.</abstract>
<identifier type="citekey">luo-etal-2026-lightmoe</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.1062/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>21124</start>
<end>21134</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T LightMoE: Task-Aware Expert Availability Management for Memory-Efficient MoE-LLM Inference
%A Luo, Puhan
%A Yao, Yunhao
%A Wang, Junyang
%A Zhang, Junyang
%A Li, Xiangyang
%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 luo-etal-2026-lightmoe
%X Mixture-of-Experts (MoE) models offer a promising path for scaling model capacity, yet their massive memory footprint poses significant challenges for deployment on resource-constrained edge devices. Existing solutions, such as static pruning or dynamic offloading, often struggle to balance model accuracy with inference latency due to irreversible information loss or prohibitive I/O overhead. In this paper, we propose LightMoE, a novel framework for memory-efficient MoE inference that exploits the inherent functional redundancy and temporal locality of expert activation. LightMoE employs a frequency-aware expert initialization strategy to retain a compact core of resident experts and introduces a similarity-based redirection mechanism to compensate for missing experts without incurring I/O costs. Furthermore, it incorporates a lightweight runtime manager that performs coarse-grained, task-level expert replacement to adapt to shifting data distributions. Empirical evaluations on representative edge platforms demonstrate that LightMoE achieves a superior accuracy-efficiency trade-off, improving average accuracy by 4.3% over static pruning and 2.4% over dynamic swapping methods, while maintaining inference latency comparable to strictly pruned models.
%U https://aclanthology.org/2026.findings-acl.1062/
%P 21124-21134
Markdown (Informal)
[LightMoE: Task-Aware Expert Availability Management for Memory-Efficient MoE-LLM Inference](https://aclanthology.org/2026.findings-acl.1062/) (Luo et al., Findings 2026)
ACL