François Portier
Associate Professor at ENSAI, member of the CREST
Associate editor for the journal Computational Statistics & Data Analysis
Head of the Master for Smart Data science
Short Bio. I did my PhD thesis in Rennes at IRMAR. I was a PostDoc at UCLouvain from 2013 to 2016. Next I moved to Télécom Paris as an assistant professor till 2021.
Research topics
Monte Carlo methods
Statistical learning
Semiparametric statistics
Lecture notes
Ordinary least squares. PDF
Monte Carlo methods. PDF
Bootstrap and resampling methods. PDF
Recent talks
SGD with coordinate sampling: Theory and Practice slides (used at Chichi León and Marc Lavielle conference in December 22)
Nearest neighbor process. slides (used at UCLouvain in April 22 and Vienna in October 22)
Control variates with nearest neighbor. slides (used at the gdr-ISIS workshop on Negative dependence in September 22)
Asymptotic Optimality with Conditioned SGD. slides (used at LAREMA in april 2021)
A guided tour in Monte Carlo. slides (used at KUL in march 2019)
New Papers
Consistent Cross Validation with stable learners. With Anass Aghbalou, Anne Sabourin. PDF
Tail inverse regression for dimension reduction with extreme response. With Anass Aghbalou, Anne Sabourin, Chen Zhou. PDF
Nearest neighbor process: weak convergence and non-asymptotic bound. PDF
Towards Asymptotic Optimality with Conditioned Stochastic Gradient Descent. With Rémi Leluc. PDF
CONTACT
francois.portier_gmail_com
ENSAI, Campus de Ker Lann
35170 Rennes
Office: 253
Papers
SGD with Coordinate Sampling: Theory and Practice. To appear in Journal of Machine Learning Research. With Rémi Leluc. PDF
A Quadrature Rule combining Control Variates and Adaptive Importance Sampling. To appear in in NeurIPS22. With Rémi Leluc, Johan Segers, Aigerim Zhuman. PDF
Conditional independence testing via weighted partial copulas and nearest neighbors. To appear in Journal of multivariate analysis. With Pascal Bianchi and Kevin Elgui. PDF
Risk bounds when learning infinitely many response functions by ordinary linear regression. Annales de l'Institut Henri Poincaré (B) Probabilités et Statistiques. With Vincent Plassier and Johan Segers. PDF
Adaptive Importance Sampling meets Mirror Descent: a Bias-variance tradeoff. AISTAT22 (oral presentation). With Anna Korba. PDF
Empirical risk minimization under random censorship: theory and practice. Journal of Machine Learning Research 2022. With Guillaime Ausset and Stéphan Clémençon. PDF
On an extension of the promotion time cure model. Annals of statistics 2022. With Jad Beyhum, Anouar El Ghouch and Ingrid Van Keilegom. PDF
Infinite-dimensional gradient-based descent for alpha-divergence minimisation. Annals of statistics 2021. With Kamelia Daudel and Randal Douc. PDF
Control variate selection for Monte Carlo integration. Statistics and computing 2021. With Rémi Leluc and Johan Segers. PDF
Safe and adaptive importance sampling: a mixture approach. Annals of statistics 2021. With Bernard Delyon. PDF
Nearest Neighbour Based Estimates of Gradients: Sharp Nonasymptotic Bounds and Applications. AISTAT 2021. With Guillaume Ausset and Stéphan Clémençon.
High dimensional regression for regenerative time-series: an application to road traffic modeling. Computational statistics and data analysis. 2021. With Mohammed Bouchouia. PDF
Parametric versus nonparametric : the fitness coefficient. Scandinavian journal of statistics. 2020. With Gildas Mazo. PDF
Monte Carlo integration with a growing number of control variates. Journal of applied probability. 2019. With Johan Segers. PDF
Rademacher complexity for Markov chains : Applications to kernel smoothing and Metropolis-Hastings. Bernoulli. 2019. With Patrice Bertail. PDF
Asymptotic optimality of adaptive importance sampling. NeurIPS. 2018. With Bernard Delyon. PDF
Integral estimation based on Markovian design. Advances in applied probability. 2018. With Romain Azais and Bernard Delyon. PDF
Beating Monte Carlo Integration: a Nonasymptotic Study of Kernel Smoothing Methods. AISTAT. 2018. With Stéphan Clémençon.
On the weak convergence of the empirical conditional copula under a simplifying assumption. Journal of multivariate analysis. 2018. With Johan Segers. PDF
Efficiency and bootstrap in the promotion time cure model. Bernoulli. 2017. With Anouar El Ghouch and Ingrid Van Keilegom. PDF
On the asymptotics of Z-estimators indexed by the objective functions. Electronic journal of statistics. 2016. PDF
An empirical process view of inverse regression. Scandinavian journal of statistics. 2016. PDF
Integral approximation by kernel smoothing. Bernoulli. 2016. With Bernard Delyon. PDF
Bootstrap testing of the rank of a matrix via least squared constrained estimation. Journal of the American statistical association. 2014. With Bernard Delyon. PDF
Optimal function: A new approach for covering the central subspace. Journal of multivariate analysis. 2013. With Bernard Delyon. PDF
PhD students
Anass Aghbalou. With Anne Sabourin.
Rémi Leluc. With Pascal Bianchi.
Guillaume Ausset. With Stéphan Clémençon. Defended in November 2021.
Kamelia Daudel. With Randal Douc and François Roueff. Defended in September 2021.
Kevin Elgui. With Pascal Bianchi. Defended in December 2020.
Other Links
Arthur Charpentier's blog about statistics, actuarial sciences and R,
Christian Robert's blog about Bayesian statistics and Monte Carlo methods,
more about the wind in Bretagne.