Gilles STOLTZ

Affiliate Professor

Economics and Decision Sciences

stoltz@hec.fr http://stoltz.perso.math.cnrs.fr/

Biography

Gilles Stoltz is an alumnus of the Ecole Normale Supérieure of Cachan and holds a Ph.D.from the University Paris-Sud Orsay since 2005.

His research interests focus on the sequential prediction of arbitrary sequences, both from an industrial viewpoint (prediction of electricity consumption or air-quality forecasting) and from a game-theoretic viewpoint (when players aim at forecasting each other's actions).

He is a researcher for CNRS at the GREGHEC laboratory. In addition, he is an affiliate professor at HEC, where he teaches the foundations of statistics in the bachelor programme.

He was awarded in 2008 the Marie-Jeanne Laurent-Duhamel prize of the young statistician by the French statistical society.

Personal web page: http://stoltz.perso.math.cnrs.fr/

Filter

Scientific articles

Explore first, exploite next: the true shape of regret in bandit problems

Mathematics of Operations Research, december, 2 2019, vol. 44, n° 2, pp 377-399, (in coll. with A. GARIVIER, P. MENARD)

Fundamentals and exchange rate forecastability with simple machine learning methods

Journal of International Money and Finance, November 2018, vol. 88, pp 1-24, (in coll. with C. AMAT, T. K. MICHALSKI)

A primal condition for approachability with partial monitoring

Journal of Dynamics and Games, July 2014, vol. 1, n° 3, pp 447-469, (in coll. with S. MANNOR, V. PERCHET)

Set-Valued Approachability and Online Learning with Partial Monitoring

Journal of Machine Learning Research, October 2014, vol. 15, pp 3247-3295, (in coll. with S. MANNOR, V. PERCHET)

Kullback-Leibler upper confidence bounds for optimal sequential allocation

Annals of Statistics, June 2013, vol. 41, n° 3, pp 1516-1541, (in coll. with O. CAPPE, A. GARIVIER, O.-A. MAILLARD, R. MUNOS)

Forecasting electricity consumption by aggregating specialized experts

Machine Learning, February 2013, vol. 90, n° 2, pp 231-260, (in coll. with M. DEVAINE, P. GAILLARD, Y. GOUDE)

Mirror descent meets fixed share (and feels no regret)

Advances in Neural Information Processing Systems, 2012, vol. 25, pp 980-988, (in coll. with N. CESA-BIANCHI, P. GAILLARD, G. LUGOSI)

X-armed bandits

Journal of Machine Learning Research, 2011, vol. 12, pp 1655-1695, (in coll. with C. Szepesvári, S. Bubeck, R. Munos)

Lipschitz Bandits without the Lipschitz Constant

Lecture Notes in Computer Science, 2011, vol. 6925, pp 144-158, (in coll. with S. Bubeck, J. Yuan Yu)

Pure exploration in finitely-armed and continuous-armed bandits

Theoretical Computer Science, April 2011, vol. 412, n° 19, pp 1832-1852, (in coll. with S. Bubeck, R. Munos)

Books

Exercices d'oral de mathématiques BL - ECE - ECS corrigés et commentés par leurs auteurs

Ellipses 2015 (in coll. with S. ARLOT, A. GARIVIER)

Statistique mathématique en action

Vuibert 2012 (in coll. with V. Rivoirard)

Statistique mathématique en action

Vuibert 2009 (in coll. with V. Rivoirard)

Proceedings

A second-order bound with excess losses

Proceedings of The 27th Conference on Learning Theory , 2014 , 35 (P. GAILLARD, T. VAN ERVEN)

Approachability in unknown games: Online learning meets multi-objective optimization

Proceedings of The 27th Conference on Learning Theory , 2014 , 35 (S. MANNOR, V. PERCHET)

A Finite-Time Analysis of Multi-armed Bandits Problems with Kullback-Leibler Divergences

24th Annual Conference on Learning Theory : COLT11 , 2011 , Budapest (O.-A. Maillard R. Munos)

Lipschitz Bandits without the Lipschitz Constant

ALT 2011 - Proceedings of the 22nd International Conference on Algorithmic Learning Theory , 2011 (S. Bubeck, J. Yu)

Working papers

Forecasting Exchange Rates Better than the Random Walk Thanks to Machine Learning Techniques

Cahier de Recherche du Groupe HEC , 2014

A finite-time analysis of multiarmed bandits problems with Kullback-Leibler divergences

Mimeo , 2011

Robust approachability and regret minimization in games with partial monitoring

Mimeo , 2011

Do countries falsify economic date strategically? Some evidence that they do

Cahier de Recherche du Groupe HEC , 2010

Forecasting the electricity consumption by aggregation of specialized experts ; application to Slovakian and French country-wide hourly predictions

Mimeo , 2010