Preprints and articles under review

1. [S3] Daniil Tiapkin, Evgenii Chzhen and Gilles Stoltz, Narrowing the gap between adversarial and stochastic MDPs via policy optimization, 2024.
    
2. [S2] Matthieu Jonckheere, Chiara Mignacco, and Gilles Stoltz, Symphony of experts: orchestration with adversarial insights in reinforcement learning, 2023.
    
3. [S1] Evgenii Chzhen, Christophe Giraud, and Gilles Stoltz, Parameter-free projected gradient descent, 2023.

Articles in proceedings of international conferences 

• [C11] Evgenii Chzhen, Christophe Giraud, Zhen Li, and Gilles Stoltz, Small total-cost constraints in contextual bandits with knapsacks, with application to fairness, Neurips, 2023.
   
• [C10] Antoine Barrier, Aurélien Garivier, and Gilles Stoltz, On best-arm identification with a fixed budget in non-parametric multi-armed bandits, ALT, 2023.
   
• [C9] Zhen Li and Gilles Stoltz, Contextual bandits with knapsacks for a conversion model, Neurips, 2022. 
   
• [C8] Evgenii Chzhen, Christophe Giraud, and Gilles Stoltz, A unified approach to fair online learning via Blackwell approachability, Neurips (with spotlight presentation), 2021. 
   
• [C7] Margaux Brégère, Yannig Goude, Pierre Gaillard, and Gilles Stoltz, Target tracking for contextual bandits: Application to demand side management, ICML, 2019.
   
• [C6] Pierre Gaillard, Sébastien Gerchinovitz, Malo Huard, and Gilles Stoltz, Uniform regret bounds over R^d for the sequential linear regression problem with the square loss, ALT, 2019.
   
• [C5] Shie Mannor, Vianney Perchet, and Gilles Stoltz, Approachability in unknown games: Online learning meets multi-objective optimization, COLT, 2014.
  The link above is to the version published in the COLT proceedings. An extended version is available (including results not cited in the COLT version).
   
• [C4] Pierre Gaillard, Gilles Stoltz, and Tim van Erven, A second-order bound with excess losses, COLT, 2014.
   
• [C3] Nicolò Cesa-Bianchi, Pierre Gaillard, Gábor Lugosi, and Gilles Stoltz, Mirror descent meets fixed share (and feels no regret), Neurips, 2012.
   
• [C2] Sébastien Bubeck, Gilles Stoltz, and Jia Yuan Yu, Lipschitz bandits without the Lipschitz constant, ALT, 2011.
   
• [C1] Gábor Lugosi, Omiros Papaspiliopoulos, and Gilles Stoltz, Online multi-task learning with hard constraints, COLT, 2009.

Extended abstracts of the articles [J1], [J2], [J5], [J6], and [J15] below were published in the Proceedings of COLT. 
An extended abstract of [J9] below was published in the Proceedings of ALT.
An extended abstract of [J10] below was published in the Proceedings of NeurIPS.
Article [J13] was obtained by merging and deeply reworking two articles published in the Proceedings of COLT in 2011 (one written by Odalric-Ambrym Maillard, Rémi Munos and myself, the other one by Aurélien Garivier and Olivier Cappé).

Articles in international journals 

• [J24] Elisabeth Gassiat and Gilles Stoltz, The van Trees inequality in the spirit of Hajek and Le Cam, Statistical Science, in press, 2024.
   
• [J23] Hédi Hadiji, Sébastien Gerchinovitz, Jean-Michel Loubes, and Gilles Stoltz, Diversity-preserving K-armed bandits, revisited, Transactions on Machine Learning Research, July 2024.
   
• [J22] Hédi Hadiji and Gilles Stoltz, Adaptation to the range in K-armed bandits, Journal of Machine Learning Research, 24(13):1−33, 2023.
   
• [J21] Rémi Coulaud, Christine Keribin, and Gilles Stoltz, Modeling dwell time in a data-rich railway environment: with operations and passenger flows data, Transportation Research Part C: Emerging Technologies, 146(103980), 2023.
   
• [J20] Aurélien Garivier, Hédi Hadiji, Pierre Ménard, and Gilles Stoltz, KL-UCB-switch: optimal regret bounds for stochastic bandits from both a distribution-dependent and a distribution-free viewpoints, Journal of Machine Learning Research, 23(179):1-66, 2022.
   
• [J19] Sébastien Gerchinovitz, Pierre Ménard and Gilles Stoltz, Fano's inequality for random variables, Statistical Science, 35(2):178-201, 2020.
   
• [J18] Raphaël Deswarte, Véronique Gervais, Gilles Stoltz, and Sébastien Da Veiga, Sequential model aggregation for production forecasting, Computational Geosciences, 23(5):1107-1124, 2019.
   
• [J17] Aurélien Garivier, Pierre Ménard and Gilles Stoltz, Explore first, exploite next: the true shape of regret in bandit problems, Mathematics of Operations Research, 44(2):377-399, 2019.
   
• [J16] Christophe Amat, Tomasz Michalski and Gilles Stoltz, Fundamentals and exchange rate forecastability with simple machine learning methods, Journal of International Money and Finance, 88:1-24, 2018.
   
• [J15] Shie Mannor, Vianney Perchet, and Gilles Stoltz, Set-valued approachability and online learning with partial monitoring, Journal of Machine Learning Research, 15(Oct):3247−3295, 2014.
   
• [J14] Shie Mannor, Vianney Perchet, and Gilles Stoltz, A primal condition for approachability with partial monitoring, Journal of Dynamics and Games, 1(3):447-469, 2014.
   
• [J13] Olivier Cappé, Aurélien Garivier, Odalric-Ambrym Maillard, Rémi Munos, and Gilles Stoltz, Kullback-Leibler upper confidence bounds for optimal sequential allocation, Annals of Statistics, 41(3):1516-1541, 2013.
   
• [J12] Tomasz Michalski and Gilles Stoltz, Do countries falsify economic data strategically? Some evidence that they might, The Review of Economics and Statistics, 95(2):591-616, 2013.
   
• [J11] Marie Devaine, Pierre Gaillard, Yannig Goude, and Gilles Stoltz, Forecasting electricity consumption by aggregating specialized experts; a review of the sequential aggregation of specialized experts, with an application to Slovakian and French country-wide one-day-ahead (half-)hourly predictions, Machine Learning, 90(2):231-260, 2013.
   
• [J10] Sébastien Bubeck, Rémi Munos, Gilles Stoltz, and Csaba Szepesvari, X-armed bandits, Journal of Machine Learning Research, 12(May):1655-1695, 2011.
   
• [J9] Sébastien Bubeck, Rémi Munos, and Gilles Stoltz, Pure exploration for multi-armed bandit problems, Theoretical Computer Science, 412:1832-1852, 2011.
   
• [J8] Shie Mannor and Gilles Stoltz, A geometric proof of calibration, Mathematics of Operations Research, 35:721-727, 2010.
   
• [J7] Vivien Mallet, Gilles Stoltz, and Boris Mauricette, Ozone ensemble forecast with machine learning algorithms, Journal of Geophysical Research, 114, D05307, 2009.
   
• [J6] Shie Mannor, Gábor Lugosi, and Gilles Stoltz, Strategies for prediction under imperfect monitoring, Mathematics of Operations Research, 33:513-528, 2008.
   
• [J5] Nicolò Cesa-Bianchi, Yishay Mansour, and Gilles Stoltz, Improved second-order bounds for prediction with expert advice, Machine Learning, 66:321-352, 2007. 
   
• [J4] Gilles Stoltz and Gábor Lugosi, Learning correlated equilibria in games with compact sets of strategies, Games and Economic Behavior, 59:187-208, 2007.
   
• [J3] Nicolò Cesa-Bianchi, Gábor Lugosi, and Gilles Stoltz, Regret minimization under partial monitoring, Mathematics of Operations Research, 31:562-580, 2006.
   
• [J2] Nicolò Cesa-Bianchi, Gábor Lugosi, and Gilles Stoltz, Minimizing regret with label-efficient prediction, IEEE Transactions on Information Theory, 51:2152-2162, 2005.
   
• [J1] Gilles Stoltz and Gábor Lugosi, Internal regret in on-line portfolio selection, Machine Learning, 59:125-159, 2005.