SiftMoE: Similarity-Aware Energy-Efficient Expert Selection for Wireless Distributed MoE Inference

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Qian, Chen, Xianhao, Huang, Kaibin
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911543275290624
author Chen, Qian
Chen, Xianhao
Huang, Kaibin
author_facet Chen, Qian
Chen, Xianhao
Huang, Kaibin
contents Mixture-of-Experts (MoE) architectures leverage sparse activation to enhance the scalability of large language models (LLMs), making them suitable for deployment in resource-constrained edge networks. However, the sheer number of experts often exceeds the memory capacity of individual edge nodes, necessitating wireless distributed MoE (WIDE) inference where experts are spread across multiple edge nodes. In this context, expert selection directly affects communication costs. Motivated by the similarity of experts, we propose SiftMoE, which judiciously selects or skips experts to strike a tradeoff between communication costs and inference accuracy. Specifically, we first establish theoretical bounds on the accuracy degradation resulting from expert replacement or skipping. Based on the bounds, we formulate an energy minimization problem for expert selection in WIDE inference subject to latency and accuracy constraints. In particular, for slow-fading channels, we derive optimal expert selection policies for both single-token decoding and multi-token prefilling. For fast-fading channels, we further extend our scheme to cope with rapidly varying channel conditions. Simulation results demonstrate that SiftMoE significantly reduces energy consumption while maintaining inference accuracy compared with conventional Top-K routing in WIDE systems.
format Preprint
id arxiv_https___arxiv_org_abs_2603_23888
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SiftMoE: Similarity-Aware Energy-Efficient Expert Selection for Wireless Distributed MoE Inference
Chen, Qian
Chen, Xianhao
Huang, Kaibin
Information Theory
Networking and Internet Architecture
Mixture-of-Experts (MoE) architectures leverage sparse activation to enhance the scalability of large language models (LLMs), making them suitable for deployment in resource-constrained edge networks. However, the sheer number of experts often exceeds the memory capacity of individual edge nodes, necessitating wireless distributed MoE (WIDE) inference where experts are spread across multiple edge nodes. In this context, expert selection directly affects communication costs. Motivated by the similarity of experts, we propose SiftMoE, which judiciously selects or skips experts to strike a tradeoff between communication costs and inference accuracy. Specifically, we first establish theoretical bounds on the accuracy degradation resulting from expert replacement or skipping. Based on the bounds, we formulate an energy minimization problem for expert selection in WIDE inference subject to latency and accuracy constraints. In particular, for slow-fading channels, we derive optimal expert selection policies for both single-token decoding and multi-token prefilling. For fast-fading channels, we further extend our scheme to cope with rapidly varying channel conditions. Simulation results demonstrate that SiftMoE significantly reduces energy consumption while maintaining inference accuracy compared with conventional Top-K routing in WIDE systems.
title SiftMoE: Similarity-Aware Energy-Efficient Expert Selection for Wireless Distributed MoE Inference
topic Information Theory
Networking and Internet Architecture
url https://arxiv.org/abs/2603.23888