Research

I work on statistical learning theory with a particular focus on online learning: sequential prediction and aggregation of experts, bandit problems (stochastic, adversarial, sleeping, dueling), online convex optimisation, nonparametric regression, and applications to reinforcement learning and demand forecasting. My PhD thesis, supervised by Yannig Goude and Gilles Stoltz, focused on prediction of individual sequences.

Google Scholar  ·  ORCID

Publications

Bandits and online decision making

• Enjoying Non-linearity in Multinomial Logistic Bandits. Pierre Boudart, Pierre Gaillard, Alessandro Rudi.
• Finally Rank-Breaking Conquers MNL Bandits: Optimal and Efficient Algorithms for MNL Assortment. Aadirupa Saha and Pierre Gaillard. ICLR, 2025.
• Logarithmic Regret for Unconstrained Submodular Maximization Stochastic Bandit. Julien Zhou, Pierre Gaillard, Thibaud Rahier, and Julyan Arbel. ALT, 2025.
• Covariance-Adaptive Least-Squares Algorithm for Stochastic Combinatorial Semi-Bandits. Julien Zhou, Pierre Gaillard, Thibaud Rahier, Houssam Zenati, and Julyan Arbel. NeurIPS, 2024.
• One Arrow, Two Kills: An Unified Framework for Achieving Optimal Regret Guarantees in Sleeping Bandits. Pierre Gaillard, Aadirupa Saha and Soham Dan. AISTATS, 2023.
• Versatile Dueling Bandits: Best-of-both-World Analyses for Online Learning from Preferences. Aadirupa Saha, Pierre Gaillard. ICML, 2022.
• Efficient Kernel UCB for Contextual Bandits. Houssam Zenati, Alberto Bietti, Eustache Diemert, Julien Mairal, Matthieu Martin, Pierre Gaillard. AISTATS, 2022.
• Dueling Bandits with Adversarial Sleeping. Aadirupa Saha, Pierre Gaillard. NeurIPS, 2021.
• Online Sign Identification: Minimization of the Number of Errors in Thresholding Bandits. Reda Ouhamma, Rémy Degenne, Vianney Perchet, and Pierre Gaillard. NeurIPS (Spotlight, <3% of submitted papers), 2021.
• Improved Sleeping Bandits with Stochastic Actions Sets and Adversarial Rewards. Aadirupa Saha, Pierre Gaillard, Michal Valko. ICML, 2020.
• Efficient online learning with Kernels for adversarial large scale problems. Rémi Jézéquel, Pierre Gaillard, Alessandro Rudi. NeurIPS, 2019.
• Target Tracking for Contextual Bandits: Application to Demand Side Management. Margaux Brégère, Pierre Gaillard, Yannig Goude, Gilles Stoltz. ICML, 2019.

Online learning and aggregation

• Optimised projection-free algorithms for online learning: construction and worst-case analysis. Julien Weibel, Pierre Gaillard, Wouter M Koolen, Adrien Taylor. AISTATS 2026.
• Efficient and Near-Optimal Online Portfolio Selection. Rémi Jézéquel, Dmitrii M. Ostrovskii, Pierre Gaillard. Mathematics of Operations Research. 2025.
• Structured prediction in online learning. Pierre Boudard, Pierre Gaillard, and Alessandro Rudi.
• Mixability made efficient: Fast online multiclass logistic regression. Rémi Jézéquel, Pierre Gaillard, and Alessandro Rudi. NeurIPS (Spotlight, <3% of submitted papers), 2021.
• Efficient improper learning for online logistic regression. Rémi Jézéquel, Pierre Gaillard, Alessandro Rudi. COLT, 2020.
• Uniform regret bounds over R^d for the sequential linear regression problem with the square loss. Pierre Gaillard, Sébastien Gerchinovitz, Malo Huard, Gilles Stoltz. ALT, 2019.
• Efficient online algorithms for fast-rate regret bounds under sparsity. Pierre Gaillard, Olivier Wintenberger. NeurIPS, 2018.
• Sparse Accelerated Exponential Weights. Pierre Gaillard, Olivier Wintenberger. AISTATS, 2017.
• Online learning and game theory. A quick overview with recent results and applications. Mathieu Faure, Pierre Gaillard, Bruno Gaujal, and Vianney Perchet. ESAIM: Proceedings and Surveys, 51, 246-271, 2015.
• A second-order bound with excess losses. Pierre Gaillard, Gilles Stoltz, Tim van Erven. COLT, 2014.
• Mirror descend meets fixed share (and feels no regret). Nicolò Cesa-Bianchi, Pierre Gaillard, Gàbor Lugosi, and Gilles Stoltz. NIPS, 2012.

Nonparametric methods

• Minimax Adaptive Online Nonparametric Regression over Besov spaces. Paul Liautaud, Pierre Gaillard, Olivier Wintenberger.
• Minimax-optimal and Locally-adaptive Online Nonparametric Regression. Paul Liautaud, Pierre Gaillard, Olivier Wintenberger. ALT, 2025.
• Adaptive approximation of monotone functions. Pierre Gaillard, Sébastien Gerchinovitz, Etienne de Montbrun.
• Online nonparametric regression with Sobolev kernels. Oleksandr Zadorozhnyi, Pierre Gaillard, Sebastien Gerschinovitz, Alessandro Rudi.
• Tight Nonparametric Convergence Rates for Stochastic Gradient Descent under the Noiseless Linear Model. Raphaël Berthier, Francis Bach, Pierre Gaillard. NeurIPS, 2020.
• Online nonparametric learning, chaining, and the role of partial feedback. Nicolò Cesa-Bianchi, Pierre Gaillard, Claudio Gentile, Sébastien Gerchinovitz. COLT, 2017.
• A Chaining Algorithm for Online Nonparametric Regression. Pierre Gaillard, Sébastien Gerchinovitz. COLT, 2015.
• A consistent deterministic regression tree for non-parametric prediction of time series. Pierre Gaillard, Paul Baudin.

Optimisation

• A Continuized View on Nesterov Acceleration for Stochastic Gradient Descent and Randomized Gossip. Mathieu Even, Raphaël Berthier, Francis Bach, Nicolas Flammarion, Pierre Gaillard, Hadrien Hendrikx, Laurent Massoulié, Adrien Taylor. NeurIPS (Outstanding paper award, <1‰ of submitted papers), 2021.
• Accelerated Gossip in Networks of Given Dimension using Jacobi Polynomial Iterations. Raphaël Berthier, Francis Bach, Pierre Gaillard, SIAM Journal on Mathematics of Data Science. Volume 2, Issue 1, 2020.

Reinforcement learning

• Online Markov Decision Processes with Terminal Law Constraints. Bianca Marin Moreno, Margaux Brégère, Pierre Gaillard, Nadia Oudjane. ALT 2026.
• Online Episodic Convex Reinforcement Learning. Bianca Marin Moreno, Khaled Eldowa, Margaux Brégère, Pierre Gaillard, and Nadia Oudjane. ICML, 2025.
• MetaCURL: Non-stationary Concave Utility Reinforcement Learning. Bianca Marin Moreno, Margaux Brégère, Pierre Gaillard, and Nadia Oudjane. NeurIPS, 2024.
• Efficient Model-Based Concave Utility Reinforcement Learning through Greedy Mirror Descent. Bianca Marin Moreno, Margaux Brégère, Pierre Gaillard, Nadia Oudjane. AISTATS, 2024.

Counterfactual and off-policy learning

• Counterfactual Learning of Stochastic Policies with Continuous Actions. Houssam Zenati, Alberto Bietti, Matthieu Martin, Eustache Diemert, Pierre Gaillard, Julien Mairal. TMLR, 2025
• Sequential Counterfactual Risk Minimization. Houssam Zenati, Eustache Diemert, Matthieu Martin, Julien Mairal, Pierre Gaillard. ICML, 2023.

Forecasting and applications

• (Online) Convex Optimization for Demand-Side Management: Application to Thermostatically Controlled Loads. Bianca Marin Moreno, Margaux Brégère, Pierre Gaillard, Nadia Oudjane. Journal of Optimization Theory and Applications. Volume 205, article number 43, 2025.
• Online Convex Optimization for Survival Analysis: An Adaptive and Stochastic Approach. Camila Fernandez, Pierre Gaillard, Joseph de Vilmarest, and Olivier Wintenberger. Statistical Papers. Volume 66, article number 86, 2025.
• Aggregation methods and comparative study in time-to-event analysis models. Camila Fernandez, Chung Shue Chen, Pierre Gaillard, and Alonso Silva. International Journal of Data Science and Analytics. 2024.
• Additive models and robust aggregation for GEFCom2014 probabilistic electric load and electricity price forecasting. Pierre Gaillard, Yannig Goude, and Raphaël Nedellec. International Journal of Forecasting, Volume 32, Issue 3, pages 1038–1050, 2016.
• Forecasting electricity consumption by aggregating experts; how to design a good set of experts. Pierre Gaillard, Yannig Goude. In A. Antoniadis et al. editors, Modeling and Stochastic Learning for Forecasting in High Dimensions, volume 217 of Lecture Notes in Statistics, pages 95–115. Springer, 2015.
• Forecasting electricity consumption by aggregating specialized experts. Marie Devaine, Pierre Gaillard, Yannig Goude, and Gilles Stoltz. Machine Learning, Volume 90, Issue 2, pages 231–260, 2013.

PhD thesis

• Contributions to online robust aggregation: work on the approximation error and on probabilistic forecasting. Applications to forecasting for energy markets. Pierre Gaillard. Université Paris-Sud 11, 2015.

Software

• Opera: Online Prediction by ExpeRt Aggregation. Pierre Gaillard, Yannig Goude. R package, 2016.
  
  • ShowHide abstract
  • CRAN
  • GitHub
  • Slides
  • Documentation
  • Source code
  Opera is an R package for prediction of time series based on online robust aggregation of a finite set of forecasts (machine learning method, statistical model, physical model, human expertise…). More formally, we consider a sequence of observation y(1),…,y(t) to be predicted element by element. At each time instance t, a finite set of experts provide prediction x(k,t) of the next observation y(t). Several methods are implemented to combine these expert forecasts according to their past performance (several loss functions are implemented to measure it). These combining methods satisfy robust finite time theoretical performance guarantees. We demonstrate on different examples from energy markets (electricity demand, electricity prices, solar and wind power time series) the interest of this approach both in terms of forecasting performance and time series analysis.

PhD Students and postdocs

• Pierre Boudart, co-advised with Alessandro Rudi (Inria), 2023-...
• Paul Liautaud, co-advised with Olivier Wintenberger, 2022-...

• Julien Zhou, industrial PhD with Criteo, co-advised with Julyan Arbel (Inria) and Thibaud Rahier (Criteo), 2022-2026
• Bianca Moreno, industrial PhD with EDF R&D, co-advised with Nadia Oudjane (EDF R&D) and Margaux Brégère (EDF R&D), 2022-2025. Now research scientist at CFM.
• Camila Fernandez, industrial PhD with Nokia Bell labs, co-advised with Olivier Wintenberger, Chung Shue Chen (Nokia Bell labs), and Alonso Silva (Nokia Bell Labs). 2020-2024.
• Houssam Zenati, industrial PhD with Criteo, co-advised with Julien Mairal and Eustach Ziemert (Criteo). 2019-2023. Now postdoc at Inria Paris-Saclay.
• Rémi Jézéquel, co-advised with Alessandro Rudi, 2019-2023. Now research scientist at CFM.
• Rémy Degenne, postdoc, january-august 2020. Now researcher at Inria Lille.
• Raphaël Berthier, co-advised with Francis Bach, 2018-2021.
• Margaux Brégère, industrial PhD with EDF R&D, co-advised with Gilles Stoltz and Yannig Goude (EDF R&D), 2017-2020. Now researcher at EDF R&D.

School

• Prévision de la consommation électrique (à court terme) par agrégation séquentielle d'experts spécialisés, 2011. Under the supervision of Yannig Goude and Gilles Stoltz
  
  • Slides (fr)
  • Technical report
• Invert sparse matrices with gaussian belief propagation algorithm, 2010. Under the supervision of Devavrat Shah
• The Lasso, or how to choose among a large set of variables with few observations, 2009. Pierre Gaillard and Anisse Ismaili. Under the supervision of Sylvain Arlot