PhD defense of Xavier CASSAGNOU

Xavier CASSAGNOU

From Koopman operators to graph neural networks: a unified framework for structured time series modeling

Abstract

The growing availability of multivariate waveform data from global monitoring networks, such as the International Monitoring System, has opened new plans for detecting and interpreting geophysical events. Despite advances, current processing systems remain limited by their reliance on rule-based methods and human expertise, particularly when waveforms propagate through changeful media such as atmosphere. This thesis introduces machine learning approaches grounded in physical principles, designed specifically for the structure and challenges of global sensor networks.

Two complementary directions are explored. The first focuses on single-station infrasound data, where we use Koopman autoencoders to model and forecast low-frequency atmospheric dynamics, separating them from signals of interest. The second targets multi-station seismic data, leveraging Graph Neural Networks to exploit spatial relationships and waveform characteristics for improved event characterization, especially in estimating magnitudes. By combining dynamical systems theory with deep learning and graph-based models, this work aims to enhance the accuracy and the interpretability of global monitoring pipelines.

Keywords

Machine learning, Time series, Koopman autoencoders, Dynamical systems, Graph neural networks,

Supervision

• Christophe MILLET
• Mathilde MOUGEOT

Jury 

• Quentin BLETERY, Directeur de recherche, IRD - Géoazur, Rapporteur
• Stéphane CHRETIEN, Professeur, Université Lumière Lyon 2, Rapporteur
• Aurélie FISCHER, Professeure, Université Paris Cité, Examinateur
• Jean-Christophe LOISEAU, Maître de conférences, École Nationale Supérieure d'Arts et Métiers, Examinateur
• Laure QUIVY, Maître de conférences, ENS Paris-Saclay, Examinateur
• Frédéric PASCAL, Professeur des universités, ENS Paris-Saclay, Examinateur