Papers

Journal Papers

[8]High Probability and Risk-Averse Guarantees for Stochastic Saddle Point Problems.
Y. Laguel, N. S. Aybat, M. Gürbüzbalaban.
Journal of Machine Learning Research.
PDF

[7] Chance constrained problems: a bilevel convex optimization perspective.
Y. Laguel, W. Van Ackooij, J. Malick.
Computational Optimization and Applications.
PDF   Code   Video

[6] Push-Pull with Device Sampling.
Y.-G. Hsieh, Y. Laguel, F. Iutzeler, J. Malick.
Transaction on Automatic Control.
PDF

[5] Federated Learning with Heterogeneous Data: A Superquantile Optimization Approach.
Y. Laguel, K. Pillutla, J. Malick, Z. Harchaoui. Machine Learning Journal.
PDF

[4] Superquantile-based learning: a direct approach using gradient-based optimization
Y. Laguel, J. Malick, Z. Harchaoui. (Journal version of 1)
Journal of Signal Processing Systems.
PDF   Code   Video

[3] Superquantiles at Work: Machine Learning Applications and Efficient (Sub)gradient Computation
Y. Laguel, K. Pillutla, J. Malick, Z. Harchaoui.
Set-Valued and Variational Analysis.
PDF

[2] On the convexity of level-sets of probability functions.
Y. Laguel, W. Van Ackooij, J. Malick, G. M. Ramalho.
Journal of Convex Analysis.
PDF

[1] Randomized progressive hedging methods for multi-stage stochastic programming.
G. Bareilles, Y. Laguel, D. Grishchenko, F. Iutzeler, J. Malick.
Annals of Operations Research.
PDF   Code

Conference Papers

[5] High-probability complexity guarantees for nonconvex minimax problems.
Y. Laguel, Y. Syed, N. S. Aybat, M. Gurbuzbalaban. NeurIPS 2024.
PDF

[4] Tackling Distribution Shifts in Federated Learning with Superquantile Aggregation.
Y. Laguel, K. Pillutla, J. Malick, Z. Harchaoui.
NeurIPS 2022 Workshop on Distribution Shifts: Connecting Methods and Applications - Spotlight Paper.
PDF

[3] Differentially Private Federated Quantiles with the Distributed Discrete Gaussian Mechanism.
Y. Laguel, K. Pillutla, J. Malick, Z. Harchaoui.
International Workshop on Federated Learning: Recent Advances and New Challenges.
PDF

[2] A Superquantile Approach for Federated Learning with Heterogeneous Devices.
Y. Laguel, K. Pillutla, J. Malick, Z. Harchaoui.
Proceedings of the 55th Annual Conference on Information Sciences and Systems (CISS 2021).
PDF   Code   Video

[1] First-order optimization for superquantile-based supervised learning.
Y. Laguel, J. Malick, Z. Harchaoui.
MLSP 2020 - Best Student Paper Award.
PDF   Code   Video

Ph.D. Dissertation

Risk-averse optimization: models, algorithms and applications in machine learning.
Y. Laguel.
PhD Dissertation.
PDF