Olivier Fercoq

Home Biography Publications Talks CV Software


Preprints

arXiv:2403.12579 The Smoothed Duality Gap as a Stopping Criterion I. Walwil and O. Fercoq, 2024

Papers and book chapters

journal paper Quadratic error bound of the smoothed gap and the restarted averaged primal-dual hybrid gradient, O. Fercoq, Open Journal of Mathematical Optimization 4(6), 2023.

conference paper Solving stochastic weak Minty variational inequalities without increasing batch size, T. Pethick, O. Fercoq, P. Latafat, P. Patrinos and V. Cevher, Proceedings of the International Conference on Learning Representations, 2023.

journal paper On the convergence of stochastic primal-dual hybrid gradient, A. Alacaoglu, O. Fercoq, and V. Cevher, SIAM Journal on Optimization 32(2), p. 1288-1318, 2022.

conference paper Escaping limit cycles: Global convergence for constrained nonconvex-nonconcave minimax problems, T. Pethick, P. Latafat, P. Patrinos, O. Fercoq and V. Cevher, In 10th International Conference on Learning Representations (ICLR), 2022.

journal paper Power-aware feature selection for optimized Analog-to-Feature converter. A. Back, P. Chollet, O. Fercoq and P. Desgreys, Microelectronics Journal, p. 105386, 2022.

journal paper Linear convergence of dual coordinate descent on non-polyhedral convex problems, O. Fercoq and I. Necoara, Mathematics of Operations Research, 2022.

journal paper Scalable Semidefinite Programming, A. Yurtsever, J.A. Tropp, O. Fercoq, M. Udell and V. Cevher, SIAM Journal on Mathematics of Data Science 3 (1), pp 171-200, 2021.

journal paper Screening Rules and its Complexity for Active Set Identification, E. Ndiaye, O. Fercoq and J. Salmon, Journal of Convex Analysis 28(4), 1053-1072, 2021.

conference paper Analog-to-feature converter optimization through power-aware feature selection, A. Back, P. Chollet, O. Fercoq and P. Desgreys. In International Conference on Analog VLSI Circuits, 2021

conference paper Random extrapolation for primal-dual coordinate descent, A. Alacaoglu, O. Fercoq and V. Cevher. In Proc. of the International Conference on Machine Learning, 2020.

conference paper Improved Optimistic Algorithms for Logistic Bandits, L. Faury, M. Abeille, C. Calauzènes and O. Fercoq. In Proc. of the International Conference on Machine Learning, 2020.

conference paper Feature selection algorithms for flexible analog-to-feature converter, A. Back, P. Chollet, O. Fercoq and P. Desgreys. In Proc of the 18th IEEE International New Circuits and Systems Conference (NEWCAS), 2020.

journal paper Restarting the accelerated coordinate descent method with a rough strong convexity estimate, O. Fercoq and Z. Qu. Computational Optimization and Applications 75(1), pp 63-91, 2020.

conference paper Stochastic Frank-Wolfe Method for Composite Convex Minimization, F. Locatello, A. Yurtsever, O. Fercoq and V. Cevher. In Proc. of the conference on Neural Information Processing Systems, 2019.

journal paper An Adaptive Primal-Dual Framework for Nonsmooth Convex Minimization, Q. Tran-Dinh, A. Alacaoglu, O. Fercoq and V. Cevher. Mathematical Programming Computation, 2019.

journal paper A generic coordinate descent solver for nonsmooth convex optimization, O. Fercoq. Optimization Methods and Software, pp 1-21, 2019.

conference paper Benchmarking GNN-CMA-ES on the BBOB noiseless testbed. L. Faury, C. Calauzènes and O. Fercoq. In Proc. of the Genetic and Evolutionary Computation Conference Companion, 2019.

conference paper Almost surely constrained convex optimization, O. Fercoq, A. Alacaoglu, I. Necoara and V. Cevher. In Proc. of the International Conference on Machine Learning, 2019.

conference paper A conditional gradient-based augmented Lagrangian framework, A. Yurtsever, O. Fercoq and V. Cevher. In Proc. of the International Conference on Machine Learning, 2019.

conference paper Safe Grid Search with Optimal Complexity, E. Ndiaye, T. Le, O. Fercoq, J. Salmon and I. Takeuchi. In Proc. of the International Conference on Machine Learning, 2019.

journal paper Adaptive restart of accelerated gradient methods under local quadratic growth condition, O. Fercoq and Z. Qu. IMA Journal of Numerical Analysis, 2019.

book chapter Smooth minimization of nonsmooth functions with parallel coordinate descent methods, O. Fercoq and P. Richtárik. In: Modeling and Optimization: Theory and Applications, edited by J. Pintér and T. Terlaky, 2019.

journal paper A Coordinate Descent Primal-Dual Algorithm with Large Step Size and Possibly Non Separable Functions, O. Fercoq and P. Bianchi, SIAM Journal on Optimization, 29(1), pp. 100-134, 2019.

conference paper Generalized Concomitant Multi-Task Lasso for Sparse Multimodal Regression , M. Massias, O. Fercoq, A. Gramfort and J. Salmon. In Proc. of the 21st International Conference on Artificial Intelligence and Statistics, 2018.

conference paper A Conditional Gradient Framework for Composite Convex Minimization with Applications to Semidefinite Programming , A. Yurtsever, O. Fercoq, F. Locatello and V. Cevher. In Proc. of the International Conference on Machine Learning, 2018.

journal paper A Smooth Primal-Dual Optimization Framework for Nonsmooth Composite Convex Minimization , Q. Tran-Dinh, O. Fercoq and V. Cevher, SIAM Journal on Optimization, 28(1), pp. 96-134, 2018.

conference paper Heteroscedastic Concomitant Lasso for sparse multimodal electromagnetic brain imaging, M. Massias, O. Fercoq, A. Gramfort and J. Salmon, Proc. of the 21st International Conference on Artificial Intelligence and Statistics, 2018.

journal paper Gap Safe screening rules for sparsity enforcing penalties, E. Ndiaye, O, Fercoq, A, Gramfort and J. Salmon. Journal of Machine Learning Research, 2017.

conference paper Smooth Primal-Dual Coordinate Descent Algorithms for Nonsmooth Convex Optimization, A. Alacaoglu, Q. Tran-Dinh, O. Fercoq and V. Cevher. In Proc. of the conference on Neural Information Processing Systems, 2017.

conference paper Data sparse nonparametric regression with ε-insensitive losses , M. Sangnier, O. Fercoq and F. d'Alché-Buc. In Proc. of the Asian Conference on Machine Learning, 192-207, 2017

conference paper Efficient Smoothed Concomitant Lasso Estimation for High Dimensional Regression, E. Ndiaye, O. Fercoq, A. Gramfort, V. Leclère and J. Salmon. In Proc of the 7th International Conference on New Computational Methods for Inverse Problems, 2017.

journal paper Physiologically structured cell population dynamic models with applications to combined drug delivery optimisation in oncology, J. Clairambault and O. Fercoq. Mathematical Modelling of Natural Phenomena, 11(6), 2016.

conference paper Using Big Steps in Coordinate Descent Primal-Dual Algorithms, P. Bianchi and O. Fercoq. In Proc. of the Conference on Decision and Control, 2016.

conference paper Joint quantile regression in vector-valued RKHSs, M. Sangnier, O. Fercoq and F. d'Alché-Buc. In Proc. of the conference on Neural Information Processing Systems, 2016

conference paper GAP Safe Screening Rules for Sparse-Group-Lasso, E. Ndiaye, O. Fercoq, A. Gramfort and J. Salmon. In Proc. of the conference on Neural Information Processing Systems, 2016

conference paper SDNA: Stochastic Dual Newton Ascent for Empirical Risk Minimization, Z. Qu, P. Richtárik, M. Takáč and O. Fercoq. In Proc. of the International Conference on Machine Learning, 2016.

journal paper Optimization in High Dimensions via accelerated, parallel and proximal coordinate descent, O. Fercoq and P. Richtárik. SIAM Review, 58(4), 739–771, 2016.

journal paper Accelerated, parallel and proximal coordinate descent, O. Fercoq and P. Richtárik. SIAM J on Optimization, 25(4), 1997–2023, 2015. This paper was awarded the second prize of the 2015 IMA Leslie Fox Prize in numerical analysis.

conference paper GAP Safe screening rules for sparse multi-task and multi-class models, E. Ndiaye, O. Fercoq, A. Gramfort and J. Salmon. In Proc. of the conference on Neural Information Processing Systems, 2015.

conference paper Mind the duality gap: safer rules for the Lasso, O. Fercoq, J. Salmon and A. Gramfort. In Proc. of the International Conference on Machine Learning, 2015.

conference paper Fast Distributed Coordinate Descent for Non-Strongly Convex Losses. O. Fercoq, Z. Qu, P. Richtárik and M. Takáč. In Proc. of the IEEE international workshop on Machine Learning for Signal Processing (MLSP), 6 pages, 2014.

journal paper Synchronisation and control of proliferation in cycling cell population models with age structure, F. Billy, J. Clairambault, O. Fercoq, S. Gaubert, T. Lepoutre, T. Ouillon and S. Saito, Mathematics and Computers in Simulation, 96, pages 66-94, 2014, doi:10.1016/j.matcom.2012.03.005

journal paper Perron vector optimization applied to search engines, O. Fercoq, Applied Numerical Mathematics, 75, pages 77-99, 2014. doi:10.1016/j.apnum.2012.12.006

journal paper Ergodic Control and Polyhedral approaches to PageRank Optimization, O. Fercoq, M. Akian, M. Bouhtou and S. Gaubert, IEEE Transactions on Automatic and Control, 58(1), pages 134-148, 2013. doi:10.1109/TAC.2012.2226103.

conference paper Parallel coordinate descent for the Adaboost problem, O. Fercoq, in Proc of the International Conference on Machine Learning and Applications (ICMLA), 2013.

conference paper PageRank optimization applied to spam detection, O. Fercoq, in 6th International conference on NETwork Games, COntrol and OPtimization (Netgcoop), 2012.

book chapter Optimisation of Cancer Drug Treatments Using Cell Population Dynamics, F. Billy, J. Clairambault and O. Fercoq, pp.257-299 in: Mathematical Methods and Models in Biomedicine, edited by U. Ledzewicz, H. Schättler, A. Friedman and E. Kashdan, Springer, 2013.

conference paper Proliferation in Cell Population Models with Age Structure, F. Billy, J. Clairambault, O. Fercoq, S. Gaubert, T. Lepoutre and T. Ouillon, Numerical Analysis and Applied Mathematics ICNAAM 2011, AIP Conf. Proc. 1389, 1212-12 15

Theses

HDR thesis

Coordinate descent methods for non-differentiable convex optimisation problems, defended on 5 December 2019, Sorbonne Université, France.

Jury: Jérôme Malick, Arkadi Nemirovski, Yurii Nesterov, Pascal Bianchi, Antonin Chambolle and Emmanuel Trélat.

PhD thesis

Optimization of Perron eigenvectors and applications: from web ranking to chronotherapeutics, 17 September 2012, Ecole Polytechnique, France. Slides

Jury: Marianne Akian (co-supervisor), Konstantin Avratchenkov (reviewer), Mustapha Bouhtou (co-supervisor), Jean Clairambault (president of the jury), Michel De Lara (examiner), Stephane Gaubert (supervisor), Roberto Tempo (examiner) and Paul Van Dooren (reviewer).

Best PhD Thesis Prize awarded by the Gaspard Monge Program for Optimization and Operations Research, sponsored by ROADEF (French Operations Research Society) and SMAI (French Society for Industrial and Applied Mathematics).

Technical reports

Rapport de chemin vers la thèse Efficiency of Primal-Dual Hybrid Gradient method, Rustem Islamov, Institut Polytechnique de Paris, 2022

Rapport de stage de M2 Lower bounds and primal-dual methods for affinely constrained convex optimization under metric subregularity, Oskar Rynkiewicz, Université Paris-Dauphine, 2020

arXiv:1901.11271 Improving Evolutionary Strategies with Generative Neural Networks, L. Faury, C. Calauzènes, O. Fercoq and S. Krichen, 2019.

arXiv:1805.08594 Neural Generative Models for Global Optimization with Gradients, L. Faury, Flavian Vasile, Clément Calauzènes, Olivier Fercoq, 2018.

arXiv:1706.05837 Smoothing technique for nonsmooth composite minimization with linear operator, Quang Van Nguyen, Olivier Fercoq, Volkan Cevher, 2017.

arXiv:1609.07358 Restarting accelerated gradient methods with a rough strong convexity estimate, O. Fercoq and Z. Qu, 2016.

Universal coordinate descent and line search for parallel coordinate descent. O. Fercoq and P. Richtárik. 2013, Available on request.

arXiv:1205.6727 Convergence of Tomlin's HOTS algorithm, O. Fercoq, 2012.

Design by Marion Chagne-Fercoq