Horseshoe Mixtures-of-Experts (HS-MoE)

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Polson, Nick, Sokolov, Vadim
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911374819459072
author Polson, Nick
Sokolov, Vadim
author_facet Polson, Nick
Sokolov, Vadim
contents Horseshoe mixtures-of-experts (HS-MoE) models provide a Bayesian framework for sparse expert selection in mixture-of-experts architectures. We combine the horseshoe prior's adaptive global-local shrinkage with input-dependent gating, yielding data-adaptive sparsity in expert usage. Our primary methodological contribution is a particle learning algorithm for sequential inference, in which the filter is propagated forward in time while tracking only sufficient statistics. We also discuss how HS-MoE relates to modern mixture-of-experts layers in large language models, which are deployed under extreme sparsity constraints (e.g., activating a small number of experts per token out of a large pool).
format Preprint
id arxiv_https___arxiv_org_abs_2601_09043
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Horseshoe Mixtures-of-Experts (HS-MoE)
Polson, Nick
Sokolov, Vadim
Machine Learning
Horseshoe mixtures-of-experts (HS-MoE) models provide a Bayesian framework for sparse expert selection in mixture-of-experts architectures. We combine the horseshoe prior's adaptive global-local shrinkage with input-dependent gating, yielding data-adaptive sparsity in expert usage. Our primary methodological contribution is a particle learning algorithm for sequential inference, in which the filter is propagated forward in time while tracking only sufficient statistics. We also discuss how HS-MoE relates to modern mixture-of-experts layers in large language models, which are deployed under extreme sparsity constraints (e.g., activating a small number of experts per token out of a large pool).
title Horseshoe Mixtures-of-Experts (HS-MoE)
topic Machine Learning
url https://arxiv.org/abs/2601.09043