Saved in:
Bibliographic Details
Main Authors: Clain, Rebecca, Montesuma, Eduardo Fernandes, Mboula, Fred Ngole
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.04324
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914533386223616
author Clain, Rebecca
Montesuma, Eduardo Fernandes
Mboula, Fred Ngole
author_facet Clain, Rebecca
Montesuma, Eduardo Fernandes
Mboula, Fred Ngole
contents Decentralized multi-source domain adaptation seeks to transfer knowledge from multiple heterogeneous and related source domains to an unlabeled target domain in a decentralized setting. We address this challenge through a fully decentralized federated approach, DeFed-GMM-DaDiL, an extension of the GMM-Dataset Dictionary Learning (DaDiL) framework. Each client models its dataset as a Gaussian Mixture Model (GMM), and the federation jointly approximates them via labeled Wasserstein barycenters of shared, learnable GMM atoms. This design enables adaptation without a central server while preserving clients' privacy. We empirically study the stability of the learned representations in scenarios where the target domain has missing classes. Empirical results demonstrate that DeFed-GMM-DaDiL maintains stable and consistent shared representations across clients, effectively reconstructs missing classes, and achieves competitive performance on multi-source domain adaptation benchmarks.
format Preprint
id arxiv_https___arxiv_org_abs_2605_04324
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle DeFed-GMM-DaDiL: A Decentralized Federated Framework for Domain Adaptation
Clain, Rebecca
Montesuma, Eduardo Fernandes
Mboula, Fred Ngole
Machine Learning
Decentralized multi-source domain adaptation seeks to transfer knowledge from multiple heterogeneous and related source domains to an unlabeled target domain in a decentralized setting. We address this challenge through a fully decentralized federated approach, DeFed-GMM-DaDiL, an extension of the GMM-Dataset Dictionary Learning (DaDiL) framework. Each client models its dataset as a Gaussian Mixture Model (GMM), and the federation jointly approximates them via labeled Wasserstein barycenters of shared, learnable GMM atoms. This design enables adaptation without a central server while preserving clients' privacy. We empirically study the stability of the learned representations in scenarios where the target domain has missing classes. Empirical results demonstrate that DeFed-GMM-DaDiL maintains stable and consistent shared representations across clients, effectively reconstructs missing classes, and achieves competitive performance on multi-source domain adaptation benchmarks.
title DeFed-GMM-DaDiL: A Decentralized Federated Framework for Domain Adaptation
topic Machine Learning
url https://arxiv.org/abs/2605.04324