MAJORIS is a scientific project funded by the European Research Council under a Starting Grant (2020-2024) and coordinated by Emilie Chouzenoux at Inria Saclay, France.

Project abstract

Mathematical optimization is the key to solving many problems in science, based on the observation that physical systems obey a general principle of least action. While some problems can be solved analytically, many more can only be solved via numerical algorithms. Research in this domain has proved essential over many years. In addition, science in general is changing. Increasingly, in biology, medicine, astronomy, chemistry, physics, large amounts of data are collected by constantly improving signal and image acquisition devices, that must be analyzed by sophisticated optimization tools. In this project, we consider handling optimization problems with large datasets. This means minimizing a cost function with a complex structure and many variables. The computational load for solving these problems is too heavy for even state-of-the-art algorithms. Thus, only relatively rudimentary data processing techniques are often employed, reducing the quality of the results and limiting the outcomes that can be achieved via these novel instruments. New algorithms must be designed with computational scalability, robustness, and versatility in mind.

In this context, Majorization-Minimization (MM) approaches have a crucial role to play. They consist of a class of efficient and effective optimization algorithms that benefit from solid theoretical foundations. The MAJORIS project aims at proposing a breakthrough in MM algorithms, so that they remain efficient when dealing with big data. Several challenging questions concerning algorithm design will be tackled. These include acceleration strategies, convergence analysis with complex costs and inexact schemes. Practical, massively parallel and distributed architecture implementations will be proposed.

Publications

International journals

Y. Huang, E. Chouzenoux, and J.-C. Pesquet. Unrolled Variational Bayesian Algorithm for Image Blind Deconvolution. IEEE Transactions on Image Processing, vol. 32, pp. 430-445, Dec. 2022. [pdf] [code]

V. Elvira and E. Chouzenoux. Graphical Inference in Linear-Gaussian State-Space Models. IEEE Transactions on Signal Processing, vol. 70, pp. 4757-4771, Sep. 2022. [pdf]

E. Chouzenoux and J.-B. Fest. SABRINA: A Stochastic Subspace Majorization-Minimization Algorithm. Journal of Optimization Theory and Applications, vol. 195, pp. 919-952, 2022. [pdf]

E. Chouzenoux, S. Martin and J.-C. Pesquet. A Local MM Subspace Method for Solving Constrained Variational Problems in Image Recovery. To appear in Journal of Mathematical Imaging and Vision, 2022. [pdf]

M. Savanier, E. Chouzenoux, J.-C. Pesquet and C. Riddel. Unmatched Preconditioning of the Proximal Gradient Algorithm. IEEE Signal Processing Letters, vol. 29, pp. 1122-1126, 2022. [pdf]

E. Chouzenoux, J.C. Pesquet, C. Riddell, M. Savanier and Y. Trousset. Convergence of Proximal Gradient Algorithm in the Presence of Adjoint Mismatch. Inverse Problems, vol. 37, no. 6, pp. 065009, 2021. [pdf]

V. Elvira and E. Chouzenoux. Optimized Population Monte Carlo. IEEE Transactions on Signal Processing, vol. 70, pp. 2489-2501, 2022. [pdf]

C. Lefort, M. Chalvidal, A. Parenté, V. Blanquet, H. Massias, L. Magnol, and E. Chouzenoux. FAMOUS: a fast instrumental and computational pipeline for multiphoton microscopy applied to 3D imaging of muscle ultrastructure. Journal of Physics D: Applied Physics, vol. 54, no. 27, pp. 274005, 2021. [pdf]

Conference proceedings

C. Rossignol, F. Sureau, E. Chouzenoux, C. Comtat, J.-C. Pesquet. A Bregman Majorization-Minimization Framework for PET Image Reconstruction. In Proceedings of the 29th IEEE International Conference on Image Processing (ICIP 2022), Bordeaux, France, 16-19 October 2022. [pdf]

T. Guilmeau, E. Chouzenoux and V. Elvira. Proximal-based adaptive simulated annealing for global optimization, In Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2022), May 2022, Singapore, China. [pdf]

M. Gharbi, E. Chouzenoux, J.-C. Pesquet and L. Duval. GPU-based Implementations of MM Algorithms. Application to Spectroscopy Signal Restoration. In Proceedings of the 29th European Signal Processing Conference (EUSIPCO 2021), Dublin, Ireland (virtual), August 23-27 2021. [pdf]

J.-B. Fest and E. Chouzenoux. Stochastic Majorize-Minimize Subspace Algorithm with Application to Binary Classification. In Proceedings of the 29th European Signal Processing Conference (EUSIPCO 2021), Dublin, Ireland (virtual), August 23-27 2021. [pdf]

S. Martin, E. Chouzenoux and J.-C. Pesquet. A Penalized Subspace Strategy for Solving Large-Scale Constrained Optimization Problems. In Proceedings of the 29th European Signal Processing Conference (EUSIPCO 2021), Dublin, Ireland (virtual), August 23-27 2021. [pdf]

T. Guilmeau, E. Chouzenoux and V. Elvira. Simulated annealing: a review and a new scheme, In Proceedings of IEEE Statistical Signal Processing Workshop (SSP 2021), Rio de Janeiro, Brazil (virtual), 11-14th July 2021. [pdf]

M. Savanier, E. Chouzenoux, J.C. Pesquet, C. Riddell and Y. Trousset. Proximal Gradient Algorithm in the Presence of Adjoint Mismatch. In Proceedings of the 28th European Signal Processing Conference (EUSIPCO 2020), January 18-22 2021. [pdf]

M. Chalvidal and E. Chouzenoux. Block Distributed 3MG Algorithm and its Application to 3D Image Restoration. In Proceedings of the 27th IEEE International Conference on Image Processing (ICIP 2020), Virtual Conference, October 25-28 2020. [pdf] [code]

E. Chouzenoux and V. Elvira. GraphEM: EM Algorithm for Blind Kalman Filtering under Graphical Sparsity Constraints>. In Proceedings of the 45th IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2020), Virtual Conference, May 4-8 2020. [pdf]

Technical reports

E. Chouzenoux, A. Contreras, J.C. Pesquet and M. Savanier. Convergence Results for Primal-Dual Algorithms in the Presence of Adjoint Mismatch, 2022. [pdf]

J.-B. Fest and E. Chouzenoux. Convergence Analysis of Block Majorize-Minimize Subspace Approaches, 2021. [pdf]

T. Guilmeau, E. Chouzenoux and V. Elvira. Regularized Rényi divergence minimization through Bregman proximal gradient algorithms, 2022. [pdf]

V. Elvira, E. Chouzenoux, O. D. Akyildiz and L. Martino. Gradient-based Adaptive Importance Samplers, 2022. [pdf]

Y. Huang, E. Chouzenoux, V. Elvira and J.-C. Pesquet. Efficient Bayes Inference in Neural Networks through Adaptive Importance Sampling, 2022. [pdf]

M. Chalvidal, E. Chouzenoux, J.-B. Fest and C. Lefort. Block Delayed Majorize-Minimize Subspace Algorithm for Large Scale Image Restoration, 2021. [pdf]

Hired researchers

Post-doctoral fellow:
Andrès Contreras
PhD students:
Jean-Baptiste Fest
Mathieu Vu
Thomas Guilmeau
Clément Cosserat
Research engineers:
Claire Rossignol

International research visitors

Prof. Tulay Adali
University of Baltimore, USA – Oct 2021 to May 2022
Dr. Audrey Repetti
Heriot-Watt University, Edinburgh, UK – April 2022 to July 2022
Dr. Giorgia Franchini
UNIMORE, Modena, Italy – September 2023
Prof. Hamid Krim
North Carolina State University, Raleigh, US – May-June 2023

Equipment

MAJORIS is equipped with a NVIDIA DGX-1 server for high performance deep learning training.

Interview of the project PI available here.