Antoine Chambaz

Bienvenue sur mon site web. Vous trouverez ici une présentation de mes travaux de recherche, avec des liens vers mes publications. Une autre page est dédiée aux enseignements.

[English version here.]

Je remercie V. Beffara pour son style css.

Recherche et publications

Je suis professeur à l'Université Paris Cité, membre du MAP5, son laboratoire de mathématiques appliquées. Depuis janvier 2024, j'en suis le directeur. Entre 2020 à 2023, j'ai dirigé la Fédération Parisienne de Modélisation Mathématique. De 2012 à 2017, j'ai été membre de Modal'X, le laboratoire de modélisation aléatoire de l'Université Paris Nanterre. J'ai dirigé le laboratoire de février 2014 à octobre 2017. Mon domaine de recherche relève de la statistique, dans ses dimensions théorique, algorithmique et applicative.

Domaines d'intérêt

  • Statistique semi-paramétrique
  • Apprentissage ciblé
  • Apprentissage par renforcement
  • Biostatistique
  • Applications en médecine, santé publique, sécurité automobile, linguistique
  • Analyse causale, analyse de l'importance de variables
  • Médecine personnalisée
  • Schémas adaptatifs pour les essais randomisés

Doctorants et post-doctorants

Prépublications soumises

Quantile Super Learning for independent and online settings with application to solar power forecasting, with H. Susmann (2023) — submitted. link

Inference in Marginal Structural Models by Automatic Targeted Bayesian and Minimum Loss-Based Estimation, with H. Susmann (2023) — submitted. link

Optimal tests of the composite null hypothesis arising in mediation analysis, with C. H. Miles (2021) — link

Publications

Probabilistic Prediction of Arrivals and Hospitalizations in Emergency Departments in Île-de-France, with H. Susmann, J. Josse, M. Wargon, P. Aegerter, and E. Barcy, to appear in International Journal of Medical Informatics (2024). link

AdaptiveConformal: An R Package for Adaptive Conformal Inference, with H. Susmann and J. Josse, published online in Computo (2024). link

Forecasting the cost of drought events in France by Super Learning from a short time series of many slightly dependent data, with G. Ecoto and A. F. Bibaut, published online in Computational Statistics (2024). link

Positivity-free Policy Learning with Observational Data, with P. Zhao, J. Josse and S. Yang, Proceedings of The 27th International Conference on Artificial Intelligence and Statistics (AISTATS, 2024). link

Optimal transport-based machine learning to match specific patterns: application to the detection of molecular regulation patterns in omics data, with T. T. Y. Nguyen, W. Harchaoui, L. Mégret, C. Mendoza, O. Bouaziz, and C. Neri, Journal of the Royal Statistical Society Series C: Applied Statistics, 73(3): 639-657 (2024). link

A Scale-Invariant Sorting Criterion to Find a Causal Order in Additive Noise Models, with A. Reisach, M. Tami, C. Seiler, S. Weichwald, NeurIPS2023. link

Personalized online ensemble machine learning with applications for dynamic data streams, with I. Malenica, R. V. Phillips, A. Hubbard , R. Pirracchio, and M. J. van der Laan, Statistics in Medicine (2023). link

Risk Minimization from Adaptively Collected Data: Guarantees for Supervised and Policy Learning, with A. Bibaut, M. Dimakopoulos, N. Kallus and M. J. van der Laan (2021) — Proceedings of NeurIPS2021 (34). link

Post-Contextual-Bandit Inference, with A. Bibaut, M. Dimakopoulos, N. Kallus and M. J. van der Laan (2021) — Proceedings of NeurIPS2021 (34). link

Generalized policy elimination: an efficient algorithm for nonparametric contextual bandits, with A. Bibaut and M. J. van der Laan, Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence (UAI), Proceedings of Machine Learning Research (PLMR), 124:1099-1108 (2020). PDF, supplementary PDF

L'échelle de la causalité: allégorie et formalisation, with V. Viallon, chapter 2 in the forthcoming book Causalité et statistique edited by F. Bertrand, G. Saporta and C. Thomas-Agnan (Éditions Technip, 2020).

Une brève introduction à l'apprentissage ciblé, with D. Benkeser, chapter 4 in the forthcoming book Causalité et statistique edited by F. Bertrand, G. Saporta and C. Thomas-Agnan (Éditions Technip, 2020).
See the companion tlride website.

A ride in targeted learning territory, with D. Benkeser, special issue Causalité of the Journal de la Société Française de Statistique, 161(1):201-286 (2020). link
See the companion tlride website.

Simpson's paradox, a tale of causality, with I. Drouet and S. Memetea, special issue Causalité of the Journal de la Société Française de Statistique, 161(1):42-66 (2020). link

Causality: a special issue of Journal de la Société Française de Statistique (editorial), with D. Benkeser and M. J. van der Laan, special issue Causalité of the Journal de la Société Française de Statistique, 161(1):1-3 (2020). link

Performance guarantees for policy learning, with A. R. Luedtke, Annales de l'Institut Henri Poincaré (B) Probabilités et Statistiques, 56(3):2162-2188 (2020). PDF

Prediction of an acute hypotensive episode during an ICU hospitalization with a super learner machine learning algorithm, with M. Cherifa, A. Blet, E. Gayat, M. Resche-Rigon and R. Pirracchio, Anesthesia & Analgesia, 130(5):1157-1166 (2020).

Biostatistics in Africa 2019: A special issue of The International Journal of Biostatistics, with A. Hubbard, A. R. Luedtke, and M. J. van der Laan, Int. J. Biostat., 15(2):1 (2019). link
Foreword to a DGIJB special issue paying tribute to Africa’s statistical and epidemiological communities. link.

Asymptotically optimal algorithms for budgeted multiple play bandits, with A. R. Luedtke and E. Kaufmann, Machine Learning, 108(11):1919-1949 (2019). link link

Big data and targeted machine learning in action to assist medical decision in the ICU: the past, the present and the future, with R. Pirracchio, J. Cohen, C. Lee, I. Malenica, M. Cannesson, M. Cohen, M. Resche-Rigon, and A. Hubbard, in Anaesthesia, Critical Care & Pain Medicine, 38(4): 377-384 (2019). link

Scalable collaborative targeted learning for high-dimensional data, with C. Ju, S. Gruber, S. Lendle, J. Franklin, R. Wyss, S. Schneeweiss and M. J. van der Laan, Stat. Methods in Med. Res., 28(2): 532-554 (2018). link

Contextual ranking by passive safety of generational classes of light vehicles, with Z. Ouni, C. Denis and C. Chauvel, J. R. Stat. Soc. Ser. C. Appl. Stat., 67(2): 395-416 (2018). link

C-TMLE for continuous tuning, with M. J. van der Laan and C. Ju, chapter 10 in Targeted Learning in Data Science, by S. Rose and M. J. van der Laan (Springer, 2018). link

Online targeted learning for time series, with M. J. van der Laan and S. Lendle, chapter 19 in Targeted Learning in Data Science, by S. Rose and M. J. van der Laan (Springer, 2018). link

Targeting a simple statistical bandit problem, with W. Zheng, chapter 24 in Targeted Learning in Data Science, by S. Rose and M. J. van der Laan (Springer, 2018). link

Targeted learning using adaptive survey sampling, with E. Joly and X. Mary, chapter 29 in Targeted Learning in Data Science, by S. Rose and M. J. van der Laan (Springer, 2018). link

Targeted sequential design for targeted learning inference of the optimal treatment rule and its mean reward, with W. Zheng and M. J. van der Laan, Ann. Statist., 45(6): 1-28 (2017). PDF
See also the companion supplemental article

Special issue on data-adaptive statistical inference, with A. Hubbard and M. J. van der Laan, Int. J. Biostat., 12 (1):1, DOI: 10.1515/ijb-2016-0033 (2016). link
Foreword to the DGIJB special issue on data-adaptive inference. link

Predicting is not explaining: targeted learning of the dative alternation, with G. Desagulier, Journal of Causal Inference, 4(1):1-30, DOI: 10.1515/jci-2014-0037 (2015). link
See also the slides of our presentation at the Language in Contrast conference.

tmle.npvi: targeted, integrative search of associations between DNA copy number and gene expression, accounting for DNA methylation, with P. Neuvial Bioinformatics, 31(18):3054-3056 (2015). link
See also the companion technical report, article and R package on CRAN.

Recension de l'ouvrage ``Big data. La révolution des données est en marche'' de V. Mayer-Schönberger et K. Cukier, with I. Drouet, Statistique et Société, 3(1):23-25 (2015). link to French text

Acceleration, due to occupational exposure, of time to onset of a disease, with C. Huber and D. Choudat, book chapter in Theory and practice of risk assessment, by C. Kitsos et al. (Springer Proceedings in Mathematics & Statistics 136, 2015).

Targeted Covariate-Adjusted Response-Adaptive LASSO-Based Randomized Controlled Trials, with M. J. van der Laan and W. Zheng, book chapter in Modern Adaptive Randomized Clinical Trials: Statistical, Operational, and Regulatory Aspects, by A. Sverdlov (CRC Press, 2015).

Causality, a trialogue, with I. Drouet, J-C. Thalabard, Journal of Causal Inference, 2(2): 201-241, DOI:10.1515/jci-2013-0024 (2014). published English version and version française

Analysis of the effect of occupational exposure to asbestos based on threshold regression modeling of case-control data, with D. Choudat, C. Huber, J-C Pairon, M. J. van der Laan, Biostatistics, 15(2): 327-340 (2014). link

Inference in targeted goup sequential covariate-adjusted randomized clinical trials, with M. J. van der Laan, Scand. J. Stat., 41(1):104-140, DOI:10.1111/sjos.12013 (2014). link

Estimation of a non-parametric variable importance measure of a continuous exposure, with P. Neuvial, M. J. van der Laan, Electron. J. Stat., 6:1059-1099 (2012). PDF.
See also the companion technical report, article, and our R package on CRAN.

Classification in postural style, with C. Denis, Ann. Appl. Stat., 6(3): 977-993 (2012). PDF

TMLE in Adaptive Group Sequential Covariate-Adjusted RCTs, with M. J. van der Laan, book chapter in Targeted Learning: Causal Inference for Observational and Experimental Data, by S. Rose and M. J. van der Laan (Springer, 2011).

Probability of Success of an In Vitro Fertilization Program, book chapter in Targeted Learning: Causal Inference for Observational and Experimental Data, by S. Rose and M. J. van der Laan (Springer, 2011).

Targeting the optimal design in randomized clinical trials with binary outcomes and no covariate: theoretical study, with M. J. van der Laan, Int. J. Biostat., 7(1), Article 10 (2011).

Targeting the optimal design in randomized clinical trials with binary outcomes and no covariate: simulation study, with M. J. van der Laan, Int. J. Biostat., 7(1), Article 11 (2011).

Deux modèles de Markov caché pour processus multiples et leur contribution à l'élaboration d'une notion de style postural, with I. Bonan, P-P. Vidal, Journal de la SFdS, 150(1): 28 pages (2009). PDF

A minimum description length approach to hidden Markov models with Poisson and Gaussian emissions. Application to order identification, with A. Garivier, E. Gassiat, J. Statist. Plann. Inference, 139(3): 962-977 (2009). link

Number of hidden states and memory: a joint order estimation problem for Markov chains with Markov regime, with C. Matias, ESAIM Probab. Stat., 13: 38-50 (2009). link

Control of neuronal persistent activity by voltage-dependent dendritic properties, with E. Idoux, D. Eugene, C. Magnani, J.A. White, L.E. Moore, J. Neurophysiol., 100: 1278-1286 (2008). link

Bounds for Bayesian order identification with application to mixtures, with J. Rousseau, Ann. Statist., 36(2): 938-962 (2008). PDF

Plica semilunaris temporal ectopia: an evidence of primary nasal pterygia traction, with E. Denion, P-H. Dalens, J. Petitbon, M. Gérard, Cornea, 26(3), pp. 769-777 (2007).

Testing the order of a model, Ann. Statist., 34(3): 1166-1203 (2006). PDF

Detecting abrupt changes in random fields, ESAIM: P&S, Novembre 2002, Vol. 6, pp. 189-209 (2002). link

Rapports techniques

Forecasting the cost of drought events in France by super learning, with G. Ecoto (2022) — link

One-step ahead sequential Super Learning from short time series of many slightly dependent data, and anticipating the cost of natural disasters, with G. Ecoto and A. F. Bibaut (2021) — See the technical report and companion R package SequentialSuperLearner.

Collaborative targeted inference from continuously indexed nuisance parameter estimators, with C. Ju and M. J. van der Laan (2018) — link

Two-stage, adaptive trial designs that modify both the population enrolled and the randomization probabilities, with B. Luber and M. Rosenblum (2017) — link

Practical targeted learning from large data sets by survey sampling, with P. Bertail and E. Joly (2016) — link

Data-adaptive Inference of the Optimal Treatment Rule and its Mean Reward. The Masked Bandit, with W. Zheng and M. J. van der Laan (2016). link

From contextual to global rankings by passive safety of generational classes of light vehicles, with Z. Ouni and C. Chauvel (2016) — link

Drawing valid targeted inference when covariate-adjusted response-adaptive RCT meets data-adaptive loss-based estimation, with an application to the LASSO, with W. Zheng and M. J. van der Laan (2015) — link

Targeted Covariate-Adjusted Response-Adaptive LASSO-Based Randomized Controlled Trials, with M. J. van der Laan and W. Zheng (2014). link

Targeted learning of the probability of success of an in vitro fertilization program controlling for time-dependent confounders, with S. Rose, J. Bouyer, M. J. van der Laan (2012). link

Estimation et test de l'ordre de lois, de l'importance de variables et de paramètres causaux; applications biomédicales, Habilitation à diriger des recherches, Université Paris Descartes (2011). PDF

Segmentation spatiale et sélection de modèle: théorie et applications statistiques, PhD thesis, Université Paris-Sud, Orsay (2003).

Packages R

  • SequentialSuperLearner (GitHub), implements the so-called overarching sequential super learning algorithm, a variant of the super learning algorithm. Designed to learn from times series, it sequentially identifies the best algorithm in a library, or the best combination of algorithms in the library, where the said library consists of several super learners, with G. Ecoto
  • tmle.npvi (CRAN), Targeted Learning of a NP Importance of a Continuous Exposure, with P. Neuvial
  • tsml.cara.rct (GitHub), Targeted Sequential Minimum Loss CARA RCT Design and Inference
  • condensier (GitHub), Non-parametric conditional density estimation with binned conditional histograms, with O. Sofrygin, F. Blaauw, and M. J. van der Laan
  • OnlineSuperLearner (GitHub), SuperLearner with online functionality for time-series analysis, with F. Blaauw

Et puis aussi

Responsabilité éditoriale

Je suis co-éditeur de l'International Journal of Biostatistics et éditeur associé du Journal of Causal Inference. Je l'ai aussi été pour le Scandinavian Journal of Statistics de 2016 à 2018.

Projets ANR BOLD et SPADRO

De septembre 2019 à février 2024, j'ai été le responsable scientifique du partenaire parisien du projet ANR BOLD (Beyond Online Learning for better Decision making) dirigé par V. Perchet.
De janvier 2014 à décembre 2018, j'ai été avec A. Garivier l'animateur principal du projet ANR SPADRO.