@inproceedings{qi-etal-2026-profiling,
title = "Profiling-Free Mixed-Precision Quantization for {M}o{E} {LLM}s via Fuzzy Rule Interpolation",
author = "Qi, Huachen and
Zhuo, Ruiyu and
Shi, Bowen and
Chang, Xiang and
Chao, Fei and
Shang, Changjing and
Shen, Qiang",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-long.982/",
doi = "10.18653/v1/2026.acl-long.982",
pages = "21484--21499",
ISBN = "979-8-89176-390-6",
abstract = "Large Language Models continue to scale in size and capability, driving substantial computational and memory demands.Mixture-of-Experts (MoE) architectures alleviate this cost by activating only a sparse subset of experts per token, enabling efficient scaling without proportional increases in inference compute.However, quantization in MoE models remains challenging due to heterogeneous sensitivity across experts and their internal linear layers.Existing mixed-precision frameworks such as Mixed-precision Quantization for MoE (MxMoE) require full quantization-loss evaluation for expert{--}layer{--}and-bit configurations, incurring prohibitive profiling cost.To address this, we propose \textbf{FRI-MxMoE}, a \textbf{profiling-free} mixed-precision quantization framework built on Fuzzy Rule Interpolation, designed as a drop-in replacement for the loss estimation component in MxMoE. By constructing a fuzzy rule base in the intra-expert layer feature space (bit-width, activation variance, parameter scale), our method predicts quantization error from only sparse samples, eliminating the need for dense profiling.Extensive experiments demonstrate that FRI-MxMoE accelerates the profiling phase by up to $15.7\times$ (on DeepSeek-V2) while achieving comparable or slightly superior zero-shot accuracy (e.g., $+1.04\%$ on DeepSeekV2-Lite) compared to the baseline.This enables continuous sensitivity modeling, preserves accuracy under mixed-precision allocation, and reduces offline computation by orders of magnitude."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="qi-etal-2026-profiling">
<titleInfo>
<title>Profiling-Free Mixed-Precision Quantization for MoE LLMs via Fuzzy Rule Interpolation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Huachen</namePart>
<namePart type="family">Qi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruiyu</namePart>
<namePart type="family">Zhuo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bowen</namePart>
<namePart type="family">Shi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiang</namePart>
<namePart type="family">Chang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fei</namePart>
<namePart type="family">Chao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Changjing</namePart>
<namePart type="family">Shang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qiang</namePart>
<namePart type="family">Shen</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>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</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">979-8-89176-390-6</identifier>
</relatedItem>
<abstract>Large Language Models continue to scale in size and capability, driving substantial computational and memory demands.Mixture-of-Experts (MoE) architectures alleviate this cost by activating only a sparse subset of experts per token, enabling efficient scaling without proportional increases in inference compute.However, quantization in MoE models remains challenging due to heterogeneous sensitivity across experts and their internal linear layers.Existing mixed-precision frameworks such as Mixed-precision Quantization for MoE (MxMoE) require full quantization-loss evaluation for expert–layer–and-bit configurations, incurring prohibitive profiling cost.To address this, we propose FRI-MxMoE, a profiling-free mixed-precision quantization framework built on Fuzzy Rule Interpolation, designed as a drop-in replacement for the loss estimation component in MxMoE. By constructing a fuzzy rule base in the intra-expert layer feature space (bit-width, activation variance, parameter scale), our method predicts quantization error from only sparse samples, eliminating the need for dense profiling.Extensive experiments demonstrate that FRI-MxMoE accelerates the profiling phase by up to 15.7\times (on DeepSeek-V2) while achieving comparable or slightly superior zero-shot accuracy (e.g., +1.04% on DeepSeekV2-Lite) compared to the baseline.This enables continuous sensitivity modeling, preserves accuracy under mixed-precision allocation, and reduces offline computation by orders of magnitude.</abstract>
<identifier type="citekey">qi-etal-2026-profiling</identifier>
<identifier type="doi">10.18653/v1/2026.acl-long.982</identifier>
<location>
<url>https://aclanthology.org/2026.acl-long.982/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>21484</start>
<end>21499</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Profiling-Free Mixed-Precision Quantization for MoE LLMs via Fuzzy Rule Interpolation
%A Qi, Huachen
%A Zhuo, Ruiyu
%A Shi, Bowen
%A Chang, Xiang
%A Chao, Fei
%A Shang, Changjing
%A Shen, Qiang
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-390-6
%F qi-etal-2026-profiling
%X Large Language Models continue to scale in size and capability, driving substantial computational and memory demands.Mixture-of-Experts (MoE) architectures alleviate this cost by activating only a sparse subset of experts per token, enabling efficient scaling without proportional increases in inference compute.However, quantization in MoE models remains challenging due to heterogeneous sensitivity across experts and their internal linear layers.Existing mixed-precision frameworks such as Mixed-precision Quantization for MoE (MxMoE) require full quantization-loss evaluation for expert–layer–and-bit configurations, incurring prohibitive profiling cost.To address this, we propose FRI-MxMoE, a profiling-free mixed-precision quantization framework built on Fuzzy Rule Interpolation, designed as a drop-in replacement for the loss estimation component in MxMoE. By constructing a fuzzy rule base in the intra-expert layer feature space (bit-width, activation variance, parameter scale), our method predicts quantization error from only sparse samples, eliminating the need for dense profiling.Extensive experiments demonstrate that FRI-MxMoE accelerates the profiling phase by up to 15.7\times (on DeepSeek-V2) while achieving comparable or slightly superior zero-shot accuracy (e.g., +1.04% on DeepSeekV2-Lite) compared to the baseline.This enables continuous sensitivity modeling, preserves accuracy under mixed-precision allocation, and reduces offline computation by orders of magnitude.
%R 10.18653/v1/2026.acl-long.982
%U https://aclanthology.org/2026.acl-long.982/
%U https://doi.org/10.18653/v1/2026.acl-long.982
%P 21484-21499
Markdown (Informal)
[Profiling-Free Mixed-Precision Quantization for MoE LLMs via Fuzzy Rule Interpolation](https://aclanthology.org/2026.acl-long.982/) (Qi et al., ACL 2026)
ACL
- Huachen Qi, Ruiyu Zhuo, Bowen Shi, Xiang Chang, Fei Chao, Changjing Shang, and Qiang Shen. 2026. Profiling-Free Mixed-Precision Quantization for MoE LLMs via Fuzzy Rule Interpolation. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 21484–21499, San Diego, California, United States. Association for Computational Linguistics.