Lénaïc Chizat

41 papers A* 12A 1Journal 28

Year	Rank	Type	Title / Venue / Authors
2026	J	`jnl`	Quantitative Convergence of Wasserstein Gradient Flows of Kernel Mean Discrepancies. CoRR Lénaïc Chizat, Maria Colombo, Roberto Colombo, Xavier Fernández-Real
2026	J	`jnl`	ResNets of All Shapes and Sizes: Convergence of Training Dynamics in the Large-scale Limit. CoRR Louis-Pierre Chaintron, Lénaïc Chizat, Javier Maass
2026	J	`jnl`	Sharper exponential convergence rates for Sinkhorn's algorithm in continuous settings. Math. Program. Lénaïc Chizat, Alex Delalande, Tomas Vaskevicius
2025	J	`jnl`	An Exponentially Converging Particle Method for the Mixed Nash Equilibrium of Continuous Games. Open J. Math. Optim. Guillaume Wang, Lénaïc Chizat
2025	J	`jnl`	Phase Diagram of Dropout for Two-Layer Neural Networks in the Mean-Field Regime. CoRR Lénaïc Chizat, Pierre Marion, Yerkin Yesbay
2025	J	`jnl`	The Hidden Width of Deep ResNets: Tight Error Bounds and Phase Diagrams. CoRR Lénaïc Chizat
2024	J	`jnl`	Annealed Sinkhorn for Optimal Transport: convergence, regularization path and debiasing. CoRR Lénaïc Chizat
2024	A*	`conf`	Deep linear networks for regression are implicitly regularized towards flat minima. NeurIPS Pierre Marion, Lénaïc Chizat
2024	J	`jnl`	Deep linear networks for regression are implicitly regularized towards flat minima. CoRR Pierre Marion, Lénaïc Chizat
2024	A*	`conf`	Mean-Field Langevin Dynamics for Signed Measures via a Bilevel Approach. NeurIPS Guillaume Wang, Alireza Mousavi Hosseini, Lénaïc Chizat
2024	J	`jnl`	On the Effect of Initialization: The Scaling Path of 2-Layer Neural Networks. J. Mach. Learn. Res. Sebastian Neumayer, Lénaïc Chizat, Michael Unser
2024	A*	`conf`	Open problem: Convergence of single-timescale mean-field Langevin descent-ascent for two-player zero-sum games. COLT Guillaume Wang, Lénaïc Chizat
2024	A*	`conf`	The Feature Speed Formula: a flexible approach to scale hyper-parameters of deep neural networks. NeurIPS Lénaïc Chizat, Praneeth Netrapalli
2024	J	`jnl`	Training Integrable Parameterizations of Deep Neural Networks in the Infinite-Width Limit. J. Mach. Learn. Res. Karl Hajjar, Lénaïc Chizat, Christophe Giraud
2023	J	`jnl`	Computational Guarantees for Doubly Entropic Wasserstein Barycenters via Damped Sinkhorn Iterations. CoRR Lénaïc Chizat, Tomas Vaskevicius
2023	A*	`conf`	Computational Guarantees for Doubly Entropic Wasserstein Barycenters. NeurIPS Tomas Vaskevicius, Lénaïc Chizat
2023	J	`jnl`	Doubly Regularized Entropic Wasserstein Barycenters. CoRR Lénaïc Chizat
2023	J	`jnl`	Local Convergence of Gradient Methods for Min-Max Games under Partial Curvature. CoRR Guillaume Wang, Lénaïc Chizat
2023	A*	`conf`	Local Convergence of Gradient Methods for Min-Max Games: Partial Curvature Generically Suffices. NeurIPS Guillaume Wang, Lénaïc Chizat
2023	J	`jnl`	On the Effect of Initialization: The Scaling Path of 2-Layer Neural Networks. CoRR Sebastian Neumayer, Lénaïc Chizat, Michael Unser
2023	J	`jnl`	Steering Deep Feature Learning with Backward Aligned Feature Updates. CoRR Lénaïc Chizat, Praneeth Netrapalli
2022	J	`jnl`	An Exponentially Converging Particle Method for the Mixed Nash Equilibrium of Continuous Games. CoRR Guillaume Wang, Lénaïc Chizat
2022	J	`jnl`	Convergence Rates of Gradient Methods for Convex Optimization in the Space of Measures. Open J. Math. Optim. Lénaïc Chizat
2022	J	`jnl`	Infinite-width limit of deep linear neural networks. CoRR Lénaïc Chizat, Maria Colombo, Xavier Fernández-Real, Alessio Figalli
2022	J	`jnl`	Mean-Field Langevin Dynamics : Exponential Convergence and Annealing. Trans. Mach. Learn. Res. Lénaïc Chizat
2022	J	`jnl`	Sparse optimization on measures with over-parameterized gradient descent. Math. Program. Lénaïc Chizat
2022	J	`jnl`	Symmetries in the dynamics of wide two-layer neural networks. CoRR Karl Hajjar, Lénaïc Chizat
2022	A*	`conf`	Trajectory Inference via Mean-field Langevin in Path Space. NeurIPS Lénaïc Chizat, Stephen Zhang, Matthieu Heitz, Geoffrey Schiebinger
2022	J	`jnl`	Trajectory Inference via Mean-field Langevin in Path Space. CoRR Lénaïc Chizat, Stephen Zhang, Matthieu Heitz, Geoffrey Schiebinger
2021	J	`jnl`	Overrelaxed Sinkhorn-Knopp Algorithm for Regularized Optimal Transport. Algorithms Alexis Thibault, Lénaïc Chizat, Charles Dossal, Nicolas Papadakis
2021	J	`jnl`	Training Integrable Parameterizations of Deep Neural Networks in the Infinite-Width Limit. CoRR Karl Hajjar, Lénaïc Chizat, Christophe Giraud
2020	A*	`conf`	Faster Wasserstein Distance Estimation with the Sinkhorn Divergence. NeurIPS Lénaïc Chizat, Pierre Roussillon, Flavien Léger, François-Xavier Vialard, Gabriel Peyré
2020	A*	`conf`	Implicit Bias of Gradient Descent for Wide Two-layer Neural Networks Trained with the Logistic Loss. COLT Lénaïc Chizat, Francis R. Bach
2020	J	`jnl`	Implicit Bias of Gradient Descent for Wide Two-layer Neural Networks Trained with the Logistic Loss. CoRR Lénaïc Chizat, Francis R. Bach
2020	A*	`conf`	Statistical and Topological Properties of Sliced Probability Divergences. NeurIPS Kimia Nadjahi, Alain Durmus, Lénaïc Chizat, Soheil Kolouri, Shahin Shahrampour, Umut Simsekli
2020	J	`jnl`	Statistical and Topological Properties of Sliced Probability Divergences. CoRR Kimia Nadjahi, Alain Durmus, Lénaïc Chizat, Soheil Kolouri, Shahin Shahrampour, Umut Simsekli
2019	J	`jnl`	HexaShrink, an exact scalable framework for hexahedral meshes with attributes and discontinuities: multiresolution rendering and storage of geoscience models. CoRR Jean-Luc Peyrot, Laurent Duval, Frédéric Payan, Lauriane Bouard, Lénaïc Chizat, Sébastien Schneider, Marc Antonini
2019	A*	`conf`	On Lazy Training in Differentiable Programming. NeurIPS Lénaïc Chizat, Edouard Oyallon, Francis R. Bach
2019	A	`conf`	Sample Complexity of Sinkhorn Divergences. AISTATS Aude Genevay, Lénaïc Chizat, Francis R. Bach, Marco Cuturi, Gabriel Peyré
2018	J	`jnl`	A Note on Lazy Training in Supervised Differentiable Programming. CoRR Lénaïc Chizat, Francis R. Bach
2018	A*	`conf`	On the Global Convergence of Gradient Descent for Over-parameterized Models using Optimal Transport. NeurIPS Lénaïc Chizat, Francis R. Bach