← samuelvaiter.com

talks & posters

selected invited talks, seminars, and conference posters.

oral presentations

Geometric and Computational Hardness of Bilevel Optimization

  • 2025-04-29 Journée MASAAI-PS, Nice.
  • 2025-04-17 MaLGa Opt Seminar, Genoa.
  • 2025-03-25 Bilevel Optimization and Hyperparameter Learning workshop, Lyon.

Successes and pitfalls of bilevel optimization in machine learning

  • 2024-12-05 Ulysseus Research Workshop on Mathematics in Machine Learning, online.
  • 2024-11-28 SophIA Summit, Sophia-Antipolis.
  • 2024-11-04 Seminar Imaging in Paris, Paris.

(Automatic) Iterative Differentiation: some old (& new) results slides notes

  • 2024-06-26 MTL Opt, MILA, Montréal.
  • 2024-03-06 Flatiron Institute, Center for Computational Mathematics, New York.
  • 2024-01-24 Kickoff seminar PEPR PDE-AI, Sorbonne Université, Paris.
  • 2023-11-06 Séminaire Parisien d'Optimisation (SPO), Sorbonne Université, Paris.
  • 2023-10-19 Journées du GdR Mathématiques de l'Optimisation, Perpignan.
  • 2023-04-03 Séminaire Pluridisciplinaire d'Optimisation Toulousain (SPOT), ENSEEIHT, Toulouse.
  • 2022-12-13 Journées SMAI-SIGMA, Sorbonne Université, Paris.

Hyper-parameters Selection by Automatic Differentiation slides video

  • 2022-06-27 Séminaire DANTE, ENS Lyon, Lyon.
  • 2021-12-02 Séminaire AOC, IRIT, Toulouse.
  • 2021-09-28 Séminaire Proba-Stat, LJAD, Nice.
  • 2021-07-09 Séminaire MAASAI, Inria, Sophia-Antipolis.
  • 2021-02-25 Séminaire Image Optimisation Probabilités, IMB, Virtual (Bordeaux).
  • 2020-11-12 Statistical Learning Seminars, Virtual.
  • 2020-01-16 Séminaire du GIPSA-Lab, Grenoble.

Graph Neural Networks on Large Random Graphs slides

  • 2025-02-07 Mathematical Imaging and Surface Processing, Mathematisches Forschungsinstitut Oberwolfach.
  • 2024-09-10 Statistical and Probabilistic Analysis of Random Networks and Processes, Nice.
  • 2022-06-28 Mathematical Methods of Modern Statistic, CIRM, Luminy.
  • 2022-06-13 Journée de la Société Française de Statistique, Université Claude Bernard, Lyon.

Dual Extrapolation for Sparse Generalized Linear Models slides

  • 2020-07-08 SIAM Conference on Imaging Sciences (SIAM IS'20), Virtual.

The Geometry of Sparse Analysis Regularization slides

  • 2019-09-18 French-German-Swiss Conference on Optimization (FGS'19), Université de Nice, Nice.

Forward Jacobian estimation: applications to risk estimation and re-fitting slides

  • 2019-07-16 International Congress on Industrial and Applied Mathematics (ICIAM'19), Universitat de València, Valencia.
  • 2016-05-24 SIAM Conference on Imaging Sciences (SIAM IS'16), Hotel Old Town, Albuquerque.

Finding optimal regularizations for low-complexity structures

  • 2019-07-09 Applied Inverse Problems (AIP'19), Université de Grenoble.

A Sharp Oracle Inequality for Graph-Slope slides

  • 2018-06-25 Journées Proba-Stat, Université de Franche-Comté, Besançon.
  • 2017-11-03 Journée de restitution Défi INFINITI, IHP, Paris.
  • 2017-06-29 Séminaire Sisyphe, ENS Lyon, Lyon.

Accelerated Alternating Descent Methods for Dykstra-like problems slides

  • 2017-09-08 Optimization 2017, Universidade de Lisboa, Portugal.

Model Selection for Low Complexity Priors slides older

  • 2018-06-07 SIAM Conference on Imaging Sciences (SIAM IS'18), Università di Bologna, Italy.
  • 2017-08-09 SPIE — Wavelets and Sparsity XVII, San Diego Convention Center, USA.
  • 2017-06-22 Information & Inference Prize Day, Stanford University, USA.
  • 2017-06-09 Journées de la SMAI, Azureva, Ronces-les-Bains.
  • 2017-01-19 Dirksen/Rauhut research group seminar, RWTH Aachen, Germany.
  • 2016-07-04 AIMS Conference on Dynamical Systems, Differential Equations and Applications, Orlando.
  • 2016-05-09 Congrès National d'Analyse Numérique, Obernai.
  • 2016-04-29 Journée de l'IMB, Université de Bourgogne, Dijon.
  • 2016-04-01 Séminaire MOD, XLIM, Université de Limoges.
  • 2016-02-09 Séminaire du CMAP, École Polytechnique, Palaiseau.
  • 2015-10-07 Journées du Programme Gaspard Monge, ENSTA, Palaiseau.
  • 2015-05-29 Applied Inverse Problems (AIP'15), Helsinki.
  • 2015-01-16 GT Signal, IRISA, Rennes.
  • 2014-11-13 GT SMATI, Telecom ParisTech, Paris.
  • 2014-10-28 New Trends in Calculus of Variations, RICAM, Linz.
  • 2014-02-03 GT Image et Statistiques, Université Paris-Dauphine, Paris.
  • 2013-10-24 GT Image, ENSICAEN, Caen.
  • 2013-10-17 GT Image, Université de Bordeaux.
  • 2013-09-06 GRETSI'13, Brest.

A First look at Proximal Methods (blackboard)

  • 2015-05-21 SMILE, École Normale Supérieure, Paris.

The Degrees of Freedom of the Group Lasso for a General Design slides

  • 2014-08-27 iTWIST, The Arsenal, Namur.
  • 2014-06-17 Curves and Surfaces 8th, Arts & Métiers ParisTech.
  • 2013-07-08 SPARS'13, EPFL, Lausanne.
  • 2013-01-23 Journée des nouveaux statisticiens, Université Paris-Dauphine, Paris.

Robust Sparse Analysis Regularization slides

  • 2012-11-20 SIGMA'12, Centre International de Rencontres Mathématiques, Marseille.
  • 2011-11-18 Journée "Adaptativité pour les images naturelles et les textures", Institut Henri Poincaré, Paris.
  • 2011-09-12 GT Image, Université Paris-Dauphine, Paris.

posters

Derivatives of Stochastic Gradient Descent in parametric optimization
NeurIPS 2024 Derivatives of Stochastic Gradient Descent in parametric optimization
One-step differentiation of iterative algorithms
NeurIPS 2023 spotlight One-step differentiation of iterative algorithms
What functions can GNNs compute on random graphs? The role of Positional Encodings
NeurIPS 2023 What functions can GNNs compute on random graphs? The role of Positional Encodings
On the Robustness of Text Vectorizers
ICML 2023 On the Robustness of Text Vectorizers
Automatic differentiation of nonsmooth iterative algorithms
NeurIPS 2022 Automatic differentiation of nonsmooth iterative algorithms
A framework for bilevel optimization that enables stochastic and global variance reduction algorithms
NeurIPS 2022 oral A framework for bilevel optimization that enables stochastic and global variance reduction algorithms
On the Universality of Graph Neural Networks on Large Random Graphs
NeurIPS 2021 On the Universality of Graph Neural Networks on Large Random Graphs
Convergence and Stability of Graph Convolutional Networks on Large Random Graphs
NeurIPS 2020 spotlight Convergence and Stability of Graph Convolutional Networks on Large Random Graphs
Exploiting regularity in sparse Generalized Linear Models
SPARS 2019 Exploiting regularity in sparse Generalized Linear Models