Yohann De Castro

My complete list of publications is available on HAL (Open Archives).

• Fast Spawn&Prune (FS&P): Global convergence of stochastic conic particle gradient descent via birth/death process.
  Y. De Castro, S. Gadat, C. Marteau.
  Under review, 2026.
  Optimization Probability PDF Code
• Gaussian random field's anisotropy using excursion sets.
  J.-M. Azaïs, Y. De Castro, F. Dalmao.
  Under review, 2025.
  Probability Geometry PDF Code arXiv
• Second Maximum of a Gaussian Random Field and Exact (t-)Spacing test.
  J.-M. Azaïs, Y. De Castro, F. Dalmao.
  Under review, 2025.
  Probability Geometry PDF Code arXiv
• Gaussian Mixture Model with unknown diagonal covariances via continuous sparse regularization.
  Y. De Castro, R. Giard, C. Marteau.
  Under review, 2025.
  Learning Optimization PDF arXiv Zenodo
• Effective regions and kernels in continuous sparse regularization.
  Y. De Castro, R. Gribonval, N. Jouvin.
  Under review, 2025.
  Learning Optimization PDF arXiv
• Node Regression on Latent Position Random Graphs via Local Averaging.
  M. Gjorgjevski, N. Keriven, S. Barthelmé, Y. De Castro.
  Journal of Machine Learning Research (JMLR), to appear.
  Graphs Optimization PDF arXiv
• Generalization Bounds of Surrogate Policies for Combinatorial Optimization Problems.
  P.-C. Aubin-Frankowski, Y. De Castro, A. Parmentier, A. Rudi.
  Under review, 2025.
  Learning Optimization PDF arXiv
• FastPart: Over-Parameterized Stochastic Gradient Descent for Sparse optimization on Measures.
  Y. De Castro, S. Gadat, C. Marteau.
  arXiv Preprint, 2023 (Revised 2025).
  Optimization Probability PDF arXiv
• Exact recovery of the support of piecewise constant images via total variation regularization.
  Y. De Castro, V. Duval, R. Petit.
  Inverse Problems, 40(10), 2024.
  Inverse Problems PDF Code arXiv
• SIGLE: valid Selective Inference procedure for Generalized Linear Lasso.
  Y. De Castro, Q. Duchemin.
  Under review, 2022.
  Learning Probability PDF Code arXiv
• Neural Networks beyond explainability: Selective inference for sequence motifs.
  Y. De Castro, A. Villié, P. Veber, L. Jacob.
  Transactions on Machine Learning Research (TMLR), 2023.
  Learning Probability PDF arXiv OpenReview
• Random Geometric Graph: Some recent developments and perspectives.
  Q. Duchemin, Y. De Castro.
  High Dimensional Probability IX, pp 347–392, 2023.
  Graphs Probability PDF arXiv
• Minimax Estimation of Partially-Observed Vector AutoRegressions.
  G. Dalle, Y. De Castro.
  Electronic Journal of Statistics, 2025.
  Graphs Probability PDF Code arXiv
• Markov Random Geometric Graph (MRGG): A Growth Model for Temporal Dynamic Networks.
  Y. De Castro, Q. Duchemin.
  Electronic Journal of Statistics, 16, 671-699, 2022.
  Graphs Probability PDF Code arXiv
• Towards off-the-grid algorithms for total variation regularized inverse problems.
  Y. De Castro, R. Petit, V. Duval.
  Journal of Mathematical Imaging and Vision (JMIV), 2022.
  Inverse Problems PDF Code arXiv
• Multiple Testing and Variable Selection along Least Angle Regression's path.
  Y. De Castro, J.-M. Azaïs.
  Information & Inference, 2022.
  Learning Probability PDF Code arXiv
• Three rates of convergence or separation via U-statistics in a dependent framework.
  Y. De Castro, Q. Duchemin, C. Lacour.
  Journal of Machine Learning Research (JMLR), 23(201), 1-59, 2022.
  Probability Learning PDF
• Concentration inequality for U-statistics of order two for uniformly ergodic Markov chains.
  Y. De Castro, Q. Duchemin, C. Lacour.
  Bernoulli Journal, 2022.
  Probability Learning PDF arXiv
• SuperMix: Sparse Regularization for Mixtures.
  Y. De Castro, S. Gadat, C. Marteau, C. Maugis-Rabusseau.
  Annals of Statistics, 49(3), 1779-1809, 2021.
  Learning Optimization PDF arXiv
• Dual optimal design and the Christoffel-Darboux polynomial.
  Y. De Castro, F. Gamboa, D. Henrion, J.-B. Lasserre.
  Optimization Letters, 15, 3-8, 2021.
  Design Optimization PDF
• Forecasting Nonnegative Time Series via Sliding Mask Method (SMM).
  Y. De Castro, L. Mencarelli.
  arXiv Preprint, 2021.
  Learning Probability PDF arXiv
• Testing Gaussian Process with Applications to Super-Resolution.
  Y. De Castro, J.-M. Azaïs, S. Mourareau.
  ACHA, 48, 445-481, 2020.
  Learning Probability PDF Code arXiv
• Latent distance estimation for random geometric graphs.
  Y. De Castro, E. Araya.
  NeurIPS, 2019.
  Geometry Probability PDF arXiv PMLR
• On Representer Theorems and Convex Regularization.
  Y. De Castro, C. Boyer, A. Chambolle, V. Duval, F. de Gournay, P. Weiss.
  SIAM Journal on Optimization, 29(2), 2019.
  Optimization Learning PDF arXiv
• Approximate Optimal Designs for Multivariate Polynomial Regression.
  Y. De Castro, F. Gamboa, D. Henrion, R. Hess, J.-B. Lasserre.
  Annals of Statistics, 47(1), 127–155, 2019.
  Algebra Optimization PDF Code arXiv
• Nonnegative Matrix Factorization with Side Information for Time Series Recovery.
  Y. De Castro, J.-M. Azaïs, Y. Goude, G. Hébrail, J. Mei.
  IEEE Trans. Knowledge and Data Engineering, 31(3), 2019.
  Learning Probability PDF arXiv
• Sparse Recovery from Extreme Eigenvalues Deviation Inequalities.
  Y. De Castro, S. Dallaporta.
  ESAIM P&S, 23, 430-463, 2019.
  Probability Learning PDF arXiv
• A short introduction to Moment-SoS hierarchies.
  Y. De Castro.
  Deuxième Congrès National de la SMF, 2018.
  Algebra Optimization PDF
• Adaptive Estimation of Nonparametric Geometric Graphs.
  Y. De Castro, C. Lacour, T. M. Pham Ngoc.
  Mathematical Statistics and Learning, 2(3/4), 2019.
  Graphs Probability PDF arXiv
• Nonnegative Matrix Factorization for Time Series Recovery From a Few Temporal Aggregates.
  J. Mei, Y. De Castro, Y. Goude, G. Hébrail.
  ICML, 2017.
  Optimization Learning PDF arXiv PMLR
• Power of the Spacing test for Least-Angle Regression.
  Y. De Castro, J.-M. Azaïs, S. Mourareau.
  Bernoulli Journal, 24(1), 2018.
  Probability Learning PDF arXiv
• Exact solutions to Super Resolution on semi-algebraic domains in higher dimensions.
  Y. De Castro, F. Gamboa, D. Henrion, J.-B. Lasserre.
  IEEE Trans. Info. Theory, 63(1), 2017.
  Algebra Optimization PDF Code arXiv
• Adapting to Unknown Noise Level in Sparse Deconvolution.
  Y. De Castro, C. Boyer, J. Salmon.
  Information & Inference, 6(3), 2017.
  Optimization Learning PDF arXiv
• Reconstructing undirected graphs from eigenspaces.
  Y. De Castro, T. Espinasse, P. Rochet.
  Journal of Machine Learning Research (JMLR), 18(51), 2017.
  Graphs Algebra PDF arXiv
• A Rice method proof of the Null-Space Property over the Grassmannian.
  Y. De Castro, J.-M. Azaïs, S. Mourareau.
  Annales de l'Institut Henri Poincaré, 53(4), 2017.
  Probability Geometry PDF arXiv
• Minimax adaptive estimation of non-parametric Hidden Markov Models.
  Y. De Castro, É. Gassiat, C. Lacour.
  Journal of Machine Learning Research (JMLR), 17(111), 2016.
  Probability Learning PDF Code arXiv
• Non-uniform spline recovery from small degree polynomial approximation.
  Y. De Castro, G. Mijoule.
  Journal of Mathematical Analysis and Applications (JMAA), Pages 971-992, 2015.
  Optimization Learning PDF arXiv
• Spike Detection from Inaccurate Samplings.
  Y. De Castro, J.-M. Azaïs, F. Gamboa.
  ACHA, 38(2), 2015.
  Optimization Learning PDF arXiv
• Randomized pick-freeze for sparse Sobol indices estimation in high dimension.
  Y. De Castro, A. Janon.
  ESAIM P&S, 19, 2015.
  Probability Learning PDF arXiv
• Optimal designs for Lasso and Dantzig selector using Expander Codes.
  Y. De Castro.
  IEEE Trans. Info. Theory, 60(11), 2014.
  Design Optimization PDF arXiv
• Exact Reconstruction using Beurling Minimal Extrapolation.
  Y. De Castro, F. Gamboa.
  Journal of Mathematical Analysis and Applications (JMAA), 395(1), 2012.
  Learning Optimization PDF arXiv

Ph.D. Students

Post-Docs

Office Hours: Monday 1pm-3pm, Room E6203.

• Machine Learning (2nd year -- Master 1)
  More detail at machine-learning-centrale.github.io .
• Convexity and Parsimony (3rd year -- Master 2)
  More detail at statistical-learning-centrale.github.io .