Preprints

Proceedings

• Jin, Liu, Rouyer, Chang, Wei and Luo, (NeurIPS 2023) No-Regret Online Reinforcement Learning with Adversarial Losses and Transitions 

• Rouyer, van der Hoeven, Cesa-Bianchi and Seldin (NeurIPS 2022), A Near-Optimal Best-of-Both-Worlds Algorithm for Online Learning with Feedback Graphs (also presented as a poster  at EWRL 2022)

• Rouyer, Seldin and Cesa-Bianchi (ICML 2021), An Algorithm for Stochastic and Adversarial Bandits with Switching Costs (also presented as a poster at NordicAI 2021)

• Rouyer and Seldin (COLT 2020): Tsallis-INF for Decoupled Exploration and Exploitation in Multi-armed Bandits (also presented as a poster at the DS3 2020 and MLSS 2020 summer schools, and won an award for the best poster at DS3 2020). 

Projects

• A Lower Bound for Adaptation of Multi-armed Bandits with Expert Advice to Easy Data

• Master Thesis: Adaptation of Exponential Weights to the Effective Range of the Losses