Ismael Castillo Professor Laboratoire de Probabilités, Statistique et Modélisation @Sorbonne Université Email: ismael.castillo 'at' upmc.fr Postal Address: LPSM UMR 8001 Case courrier 188 4, place Jussieu 75252 Paris Cedex 05 France

Preprints

• Deep horseshoe Gaussian processes, [PDF].
  Arxiv e-prints 2403.01737, 46p., with Thibault Randrianarisoa.
  
  
• Bayesian nonparametric statistics: St-Flour lecture notes, [PDF].
  Arxiv e-prints 2402.16422, 186p.
  
  
• Heavy-tailed Bayesian nonparametric adaptation [PDF].
  Arxiv e-prints 2308.04916, 52p., with Sergios Agapiou.
  
  
• Sharp multiple testing boundary for sparse sequences [PDF].
  Arxiv e-prints 2109.13601, 86p., with Kweku Abraham and Étienne Roquain.
  

Publications

• Bayesian Multiscale Analysis of the Cox model [PDF].
  Bernoulli, 30 (2024), no. 2, 1525-1554, with Bo Ning.
  
  
• Semiparametric inference with fractional posteriors [PDF].
  Journal of Machine Learning Research, 24, (389) 1-61 (2023), with Alice l'Huillier, Luke Travis and Kolyan Ray.
  
  
• Optional Pólya trees: posterior rates and uncertainty quantification [PDF].
  Electronic Journal of Statistics, 16 (2022), no. 2, 6267-6312, with Thibault Randrianarisoa.
  
  
• Uniform estimation in stochastic block models is slow [PDF].
  Electronic Journal of Statistics, 16 (2022), no. 1, 2947-3000, with Peter Orbanz.
  
  
• Multiple Testing in Nonparametric Hidden Markov Models: An Empirical Bayes Approach [PDF].
  Journal of Machine Learning Research, 23, (94) 1-57 (2022), with Kweku Abraham and Elisabeth Gassiat.
  
  
• Empirical Bayes cumulative ℓ-value multiple testing procedure for sparse sequences [PDF] .
  Electronic Journal of Statistics, 16 (2022), no. 1, 2033-2081, with Kweku Abraham and Étienne Roquain.
  
  
• Multiscale Bayesian survival analysis [PDF] .
  Annals of Statistics, 49 (2021), no. 6, 3559-3582, with Stéphanie van der Pas. R package BayesSurvival, vignette.
  
  
• Uncertainty quantification for Bayesian CART [PDF] .
  Annals of Statistics, 49 (2021), no. 6, 3482-3509, with Veronika Ročková.
  
  
• Bayesian inference in high-dimensional models [PDF]
  Springer volume on Data Science, (2021), to appear, with Sayantan Banerjee and Subhashis Ghosal.
  
  
• Spike and Slab Pólya tree posterior distributions: adaptive inference [PDF] . [Won the best paper prize 2021]
  Annales de l'Institut Henri Poincaré, 57 (2021), no. 3, 1521-1548, with Romain Mismer. [AIHP]
  
  
• On spike and slab empirical Bayes multiple testing [PDF] .
  Annals of Statistics, 48 (2020), no. 5, 2548-2574, with Étienne Roquain.
  
  
• Spike and slab empirical Bayes sparse credible sets [PDF] .
  Bernoulli, 26 (2020), no. 1, 127-158, with Botond Szabó.
  
  
• Empirical Bayes analysis of spike and slab posterior distributions [PDF] .
  Electronic Journal of Statistics, 12 (2018), no. 2, 3953-4001, with Romain Mismer.
  
  
• Pólya tree posterior distributions on densities [PDF] ,
  Annales de l'Institut Henri Poincaré, 53 (2017) , No. 4, 2074-2102. [AIHP]
  


• Discussion of Uncertainty Quantification for the Horseshoe by van der Pas et al. [PDF] ,
  Bayesian analysis, 12 (2017), no. 4, 1250-1253.
  


• Discussion of Sparse graphs using exchangeable random measures by Caron and Fox. [PDF] ,
  Journal of the Royal Statistical Society Series B, 79 (2017), no. 5, 1347-1348, with Tabea Rebafka.
  


• A Bernstein-von Mises theorem for smooth functionals in semiparametric models [PDF] ,
  Annals of Statistics, 43 (2015), no. 6, 2353-2383, with Judith Rousseau.
  


• Bayesian linear regression with sparse priors [PDF] ,
  Annals of Statistics, 43 (2015), no. 5, 1986-2018, with Johannes Schmidt-Hieber and Aad van der Vaart.
  


• Discussion of Bayesian nonparametric confidence sets by Szabó et al. [PDF]
  Annals of Statistics, 43 (2015), no. 4, 1437-1443.
  


• Bayesian nonparametrics, convergence and limiting shape of posterior distributions [PDF]
  Habilitation manuscript, Université Paris Diderot, 2014.
  


• On the Bernstein-von Mises phenomenon for nonparametric Bayes procedures [PDF]
  Annals of Statistics, 42 (2014), no. 5, 1941-1969, with Richard Nickl.
  


• On Bayesian supremum norm contraction rates [PDF]
  Annals of Statistics, 42 (2014), no. 5, 2058-2091.
  


• Thomas Bayes' walk on manifolds [PDF]
  Probability Theory and Related Fields, 158 (2014), no. 3-4, 665-710, with Gérard Kerkyacharian and Dominique Picard.
  


• Nonparametric Bernstein-von Mises Theorems in Gaussian white noise [PDF]
  Annals of Statistics, 41 (2013), no. 4, 1999-2028, with Richard Nickl.
  


• Needles and Straw in a Haystack: posterior concentration for possibly sparse sequences [PDF], [PDF]
  Annals of Statistics, 40 (2012), no. 4, 2069-2101. [AOS] with Aad van der Vaart.
  


• Semiparametric Bernstein-von Mises theorem and bias, illustrated with Gaussian process priors [PDF]
  Sankhya A, 74 (2012), 194-221. [SAN]
  


• A semi-parametric Bernstein-von Mises theorem for Gaussian process priors [PDF]
  Probability Theory and Related Fields, 152 (2012), no. 1, 53-99.
  


• Semiparametric shift and period estimation based on the cumulated periodogram for non-regular functions [PDF]
  Electronic Journal of Statistics 5 (2011), 102-126. [EJS] , with Eric Cator.
  


• Estimation of the law of random shift deformation [PDF]
  Mathematical Methods of Statistics, 18 (2009), no. 1, 21-42. [MMS], with Jean-Michel Loubes.
  


• Lower bounds for posterior rates with Gaussian process priors [PDF]
  Electronic Journal of Statistics 2 (2008), 1281-1299. [EJS]
  


• Semiparametric second order efficient estimation of the period of a signal [PDF]
  Bernoulli 13 (2007), no. 4, 910-932. [JBS]
  


• Exact adaptive estimation of the shape of a periodic function with unknown period corrupted by white noise [PDF]
  Mathematical Methods of Statistics 15 (2006), no. 2, 146-175. [MMS], with Céline Lévy-Leduc and Catherine Matias.
  


• Penalized profile likelihood methods and second order properties in semiparametrics
  PhD thesis, Université Paris-Sud, 2006.
  

Students

• PhD students (main supervisor)

  Paul Egels (PhD 2023-), Alice l'Huillier (PhD 2021-, CNRS-Imperial College grant), Thibault Randrianarisoa (defended 2022), Romain Mismer (defended 2019).


• Co-supervision

  Luke Travis (PhD 2021-, CNRS-Imperial College grant, main supervisor Kolyan Ray).



• Postdocs


•   Bo Ning (FSMP postdoc 2020-2021), Kweku Abraham (Hadamard postdoc, with E. Gassiat 2019-2021).

Teaching

• Second semester



•   Sorbonne Université: Cours M2 ``Bayésien non-paramétrique et applications" . Page web du cours.



•   Sorbonne Université: Cours M2 ``Topics in modern machine learning" , with Eddie Aamari, Claire Boyer and Étienne Roquain.




• First semester



•   Sorbonne Université: Cours M2 ``Apprentissage statistique" , joint course with Erwan Scornet.



•   Sorbonne Université: Cours M2 ``Statistiques non-paramétriques - Partie II" , joint course with Charlotte Dion.

Short Bio

• 2015 --           Professor of Statistics at Sorbonne Université, LPSM


• 2018 - 2023   Institut Universitaire de France, Junior Fellow


• 2010 - 2015   CNRS Researcher at LPMA


• Nov. -- 2014   Habilitation à diriger des recherches, Université Paris Diderot


• 2006 - 2009   Postdoc at the Free University of Amsterdam, supervisor Aad van der Vaart


• 2003 - 2006   PhD in Statistics at Université Paris-Sud, supervisor Elisabeth Gassiat


• 2000 - 2003   Ecole Normale Supérieure de Lyon