FedSPDnet: Geometry-Aware Federated Deep Learning with SPDnet

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Pautrel, Thibault, Bouchard, Florent, Mian, Ammar, Ginolhac, Guillaume
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913059100950528
author Pautrel, Thibault
Bouchard, Florent
Mian, Ammar
Ginolhac, Guillaume
author_facet Pautrel, Thibault
Bouchard, Florent
Mian, Ammar
Ginolhac, Guillaume
contents We introduce two federated learning frameworks for the classical SPDnet model operating on symmetric positive definite (SPD) matrices with Stiefel-constrained parameters. Unlike standard Euclidean averaging, which violates orthogonality, our approach preserves geometric structure through two efficient aggregation strategies: ProjAvg, projecting arithmetic means onto the Stiefel manifold, and RLAvg, approximating tangent-space averaging via retractions and liftings. Both methods are computationally efficient, independent of the optimizer, and enable scalable federated learning for signal processing applications whose features are SPD matrices. Simulations on EEG motor imagery benchmarks show that FedSPDnet outperforms federated EEGnet in F1 score and robustness to federation and partial participation, while using fewer parameters per communication round.
format Preprint
id arxiv_https___arxiv_org_abs_2604_22494
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle FedSPDnet: Geometry-Aware Federated Deep Learning with SPDnet
Pautrel, Thibault
Bouchard, Florent
Mian, Ammar
Ginolhac, Guillaume
Machine Learning
We introduce two federated learning frameworks for the classical SPDnet model operating on symmetric positive definite (SPD) matrices with Stiefel-constrained parameters. Unlike standard Euclidean averaging, which violates orthogonality, our approach preserves geometric structure through two efficient aggregation strategies: ProjAvg, projecting arithmetic means onto the Stiefel manifold, and RLAvg, approximating tangent-space averaging via retractions and liftings. Both methods are computationally efficient, independent of the optimizer, and enable scalable federated learning for signal processing applications whose features are SPD matrices. Simulations on EEG motor imagery benchmarks show that FedSPDnet outperforms federated EEGnet in F1 score and robustness to federation and partial participation, while using fewer parameters per communication round.
title FedSPDnet: Geometry-Aware Federated Deep Learning with SPDnet
topic Machine Learning
url https://arxiv.org/abs/2604.22494