Estelle Massart

Short biography

I graduated as an engineer in applied mathematics at UCLouvain, Belgium, in 2015, and received a PhD in mathematical engineering from UCLouvain in 2019, under the supervision of Julien Hendrickx and Pierre-Antoine Absil. My PhD thesis, entitled Data fitting on positive semidefinite matrix manifolds, was awarded the XXI Alston Householder prize. From 2020 to 2022, I was a postdoctoral researcher with the Mathematical Institute, at the University of Oxford, where I was funded by the National Physical Laboratory. From January to August 2022, I was an FNRS scientific collaborator at UCLouvain, in the mathematical engineering department, a subdivision of the ICTEAM institute. From September 2022, I am an academic faculty member of UCLouvain, working in the mathematical engineering department.

Research interests

• Machine learning/deep learning
• Optimization
• Riemannian geometry
• Applications of mathematics to sciences (special interest in biomedical applications)

PhD students

• Bastien Massion (2022-...), INMA-ICTEAM, UCLouvain.
• Timothé Taminiau (2023-...), INMA-ICTEAM, UCLouvain.
• David Thorsteinsson (2024-...), NUMA, KULeuven.
• Jana Jovcheva (2024-...), INMA-ICTEAM, UCLouvain.

Teaching

• LEPL1101, Algèbre
• LINFO1113, Algorithmique Numérique
• LEPL1507 Projet en mathématiques appliquées
• LINMA2120 Applied mathematics seminar
• LSINC1313 Algorithmique Numérique
• LINMA2361 Nonlinear dynamical systems

Co-supervised MSc theses:

• Numerical investigation of a continuous relaxation for the Column Subset Selection Problem, A. Victor, 2024
• On the training of shallow neural networks with the gradient method, Q. Brisbois, 2024
• Interpretability of Deep Unsupervised Domain Adaptation, G. Dony, UCLouvain, 2024
• River Bed Profile Evolution Subject to Stochastic Forcing by Migrating Tributaries, C. Chevillard, 2024
• Numerical comparison of MCMC methods for Quantum Tomography, D. Mokeev, 2024
• Classification of covariance matrices for EEG: how to handle the low-rank case? D. Gailly, 2022
• Accelerated First Order Methods for Non-convex Optimisation , M. Yu (University of Oxford), 2021
• Active Subspace Methods for Global Optimization , X. Liang (University of Oxford), 2021
• Positive-semidefinite matrix processing for EEG decoding and/or community detection, P. Veldeman, 2021
• Classification of biologic signals on the symmetric positive definite manifold C. Vaes, 2020
• Stochastic gradient methods for matrix completion, G. M. Bengoechea, 2018

Publications

Theses

• E. Massart, Data fitting on positive-semidefinite matrix manifolds, PhD dissertation, UCLouvain, 2019.
• E. Massart, Means and consensus on manifolds , MSc dissertation, UCLouvain, 2015.

Preprints

• C. Cartis, X. Liang, E. Massart, A. Otemissov, Learning the subspace of variation for global optimization of functions with low effective dimension, 2024. [Preprint]

Journal papers

• C. Cartis, E. Massart, A. Otemissov, Global optimization using random embeddings, Mathematical Programming, 2022. [Preprint] [Publisher's version]
• C. Cartis, E. Massart, A. Otemissov, Bound-constrained global optimization of functions with low effective dimensionality using multiple random embeddings, Mathematical Programming, 2022. [Preprint] [Publisher's version]
• A. Musolas, E. Massart, J. M. Hendrickx, P.-A. Absil, Y. Marzouk Low-rank multi-parametric covariance estimation, BIT Numerical Mathematics 62, 221–249, 2022. [Preprint] [Publisher's version]
• E. Massart, P.-A. Absil, Quotient geometry with simple geodesics for the manifold of fixed-rank positive-semidefinite matrices, SIAM Journal on Matrix Analysis and Applications 41(1), 171-198, 2020. [Preprint] [Publisher's version] Code available in the Manopt toolbox ( symfixedrankYYfactory.m )
• P.-Y. Gousenbourger, E. Massart, P.-A. Absil, Data fitting on manifolds with composite Bézier-like curves and blended cubic splines, Journal of Mathematical Imaging and Vision, 61(5), 645-671, 2019. [Preprint] [Publisher's version]
• E. M. Massart, J. M. Hendrickx, P.-A. Absil, Matrix geometric means based on shuffled inductive sequences, Linear Algebra and its Applications, 252, 334-359, 2018. [Preprint] [Publisher's version] [Code]

Conference papers

• E. Massart, V. Abrol, Coordinate descent on the Stiefel manifold for deep neural network training, ESANN 2023. [Publisher version]
• E. Massart, Improving weight clipping in Wasserstein GANs, ICPR 2022. [Preprint]
• E. Massart, Orthogonal regularizers in deep learning: how to handle rectangular matrices?, ICPR 2022. [Preprint]
• E. Massart, V. Abrol, Coordinate Descent on the Orthogonal Group for Recurrent Neural Network Training, To appear in the Proceedings of the 36th AAAI conference on artificial intelligence, 2022. [Preprint]
• C. Cartis, E. Massart, A. Otemissov, Dimensionality reduction techniques for global optimization of functions with low effective dimensionality, ICML workshop “Beyond first order methods in ML systems”, 2020. [Paper] [3 minutes presentation video]
• N.T. Son, P.-Y. Gousenbourger, E. Massart, and T. Stykel, Balanced Truncation for Parametric Linear Systems Using Interpolation of Gramians: A Comparison of Algebraic and Geometric Approaches, Model Reduction of Complex Dynamical Systems, P. Benner, T. Breiten, H. Faßbender, M. Hinze, T. Stykel and R. Zimmermann Eds, Springer Nature Switzerland - Birkhäuser, 31-51, 2022. [Preprint] [Publisher's version]
• B. Szczapa, M. Daoudi, S. Berretti, A. Del Bimbo, P. Pala, E. Massart, Fitting, Comparison, and Alignment of Trajectories on Positive Semi-Definite Matrices with Application to Action Recognition, ICCV Human Behavior Understanding workshop, 2019. [Preprint]
• E. Massart, J. M. Hendrickx, P.-A. Absil, Curvature of the manifold of fixed-rank positive-semidefinite matrices endowed with the Bures-Wasserstein metric, Geometric Science of Information: Fourth International Conference (GSI), 2019. [Preprint] [Publisher's version]
• N. Thanh Son, P.-Y. Gousenbourger, E. Massart, P.-A. Absil, Online balanced truncation for linear time-varying systems using continuously differentiable interpolation on Grassmann manifold, Proceedings of the 6th International Conference on Control, Decision and Information Technologies (CoDIT), 2019. [Preprint] [Publisher's version]
• E. Massart, P.-Y. Gousenbourger, N. Thanh Son, T. Stykel, P.-A. Absil, Interpolation on the manifold of fixed-rank positive-semidefinite matrices for parametric model order reduction: preliminary results, Proceedings of the 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN), pages 281-286, 2019. [Publisher's version]
• E. Massart, S. Chevallier, Inductive means and sequences applied to online classification of EEG, Geometric Science of Information: Third International Conference (GSI), 2017. [Preprint] [Publisher's version] [Talk]
• P.-Y. Gousenbourger, E. Massart, A. Musolas, P.-A. Absil, J. M. Hendrickx, L. Jacques, Y. Marzouk, Piecewise-Bezier C1 smoothing on manifolds with application to wind field estimation, Proceedings of the 25th European Symposium on Artifical Neural Networks, Computational Intelligence and Machine Learning (ESANN), pages 305-310, 2017. [Publisher's version]
• E. M. Massart, J. M. Hendrickx, P.-A. Absil, Extending a two-variable mean to a multi-variable mean, 24th European Symposium on Artifical Neural Networks, Computational Intelligence and Machine Learning (ESANN), Bruges, Belgium, 2016. [Publisher's version] [Code]