Saved in:
Bibliographic Details
Main Authors: Seraphim, Mathieu, Lechervy, Alexis, Yger, Florian, Brun, Luc, Etard, Olivier
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2309.07579
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929361236525056
author Seraphim, Mathieu
Lechervy, Alexis
Yger, Florian
Brun, Luc
Etard, Olivier
author_facet Seraphim, Mathieu
Lechervy, Alexis
Yger, Florian
Brun, Luc
Etard, Olivier
contents In recent years, Transformer-based auto-attention mechanisms have been successfully applied to the analysis of a variety of context-reliant data types, from texts to images and beyond, including data from non-Euclidean geometries. In this paper, we present such a mechanism, designed to classify sequences of Symmetric Positive Definite matrices while preserving their Riemannian geometry throughout the analysis. We apply our method to automatic sleep staging on timeseries of EEG-derived covariance matrices from a standard dataset, obtaining high levels of stage-wise performance.
format Preprint
id arxiv_https___arxiv_org_abs_2309_07579
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Structure-Preserving Transformers for Sequences of SPD Matrices
Seraphim, Mathieu
Lechervy, Alexis
Yger, Florian
Brun, Luc
Etard, Olivier
Machine Learning
Signal Processing
In recent years, Transformer-based auto-attention mechanisms have been successfully applied to the analysis of a variety of context-reliant data types, from texts to images and beyond, including data from non-Euclidean geometries. In this paper, we present such a mechanism, designed to classify sequences of Symmetric Positive Definite matrices while preserving their Riemannian geometry throughout the analysis. We apply our method to automatic sleep staging on timeseries of EEG-derived covariance matrices from a standard dataset, obtaining high levels of stage-wise performance.
title Structure-Preserving Transformers for Sequences of SPD Matrices
topic Machine Learning
Signal Processing
url https://arxiv.org/abs/2309.07579