Maximum Score Routing For Mixture-of-Experts

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Dong, Bowen, Fan, Yilong, Sun, Yutao, Li, Zhenyu, Pan, Tengyu, Zhou, Xun, Wang, Jianyong
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915449292193792
author Dong, Bowen
Fan, Yilong
Sun, Yutao
Li, Zhenyu
Pan, Tengyu
Zhou, Xun
Wang, Jianyong
author_facet Dong, Bowen
Fan, Yilong
Sun, Yutao
Li, Zhenyu
Pan, Tengyu
Zhou, Xun
Wang, Jianyong
contents Routing networks in sparsely activated mixture-of-experts (MoE) dynamically allocate input tokens to top-k experts through differentiable sparse transformations, enabling scalable model capacity while preserving computational efficiency. Traditional MoE networks impose an expert capacity constraint to ensure GPU-friendly computation. However, this leads to token dropping when capacity is saturated and results in low hardware efficiency due to padding in underutilized experts. Removing the capacity constraint, in turn, compromises load balancing and computational efficiency. To address these issues, we propose Maximum Score Routing ($\mathbf{MaxScore}$), a novel MoE routing paradigm that models routing as a minimum-cost maximum-flow problem and integrates a SoftTopk operator. MaxScore resolves the fundamental limitations of iterative rerouting and optimal transport formulations, achieving lower training losses and higher evaluation scores at equivalent FLOPs compared to both constrained and unconstrained baselines. Implementation details and experimental configurations can be obtained from $\href{https://github.com/dongbw18/MaxScore.git}{MaxScore}$.
format Preprint
id arxiv_https___arxiv_org_abs_2508_12801
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Maximum Score Routing For Mixture-of-Experts
Dong, Bowen
Fan, Yilong
Sun, Yutao
Li, Zhenyu
Pan, Tengyu
Zhou, Xun
Wang, Jianyong
Machine Learning
Computation and Language
Routing networks in sparsely activated mixture-of-experts (MoE) dynamically allocate input tokens to top-k experts through differentiable sparse transformations, enabling scalable model capacity while preserving computational efficiency. Traditional MoE networks impose an expert capacity constraint to ensure GPU-friendly computation. However, this leads to token dropping when capacity is saturated and results in low hardware efficiency due to padding in underutilized experts. Removing the capacity constraint, in turn, compromises load balancing and computational efficiency. To address these issues, we propose Maximum Score Routing ($\mathbf{MaxScore}$), a novel MoE routing paradigm that models routing as a minimum-cost maximum-flow problem and integrates a SoftTopk operator. MaxScore resolves the fundamental limitations of iterative rerouting and optimal transport formulations, achieving lower training losses and higher evaluation scores at equivalent FLOPs compared to both constrained and unconstrained baselines. Implementation details and experimental configurations can be obtained from $\href{https://github.com/dongbw18/MaxScore.git}{MaxScore}$.
title Maximum Score Routing For Mixture-of-Experts
topic Machine Learning
Computation and Language
url https://arxiv.org/abs/2508.12801