Csaba Szepesvári

Csaba Szepesvári is a Professor in the Department of Computing Science at the University of Alberta and a Senior Staff Research Scientist at DeepMind, leading the Foundations team. He holds the Canada CIFAR AI Chair and is a Fellow at Amii. His research focuses on reinforcement learning (RL), online learning, and theoretical foundations of sequential decision-making. Key areas include exploration-exploitation strategies, Markov Decision Processes, and stochastic control. He has co-authored over 225 publications, including seminal works like Bandit Algorithms (2020) and Algorithms for Reinforcement Learning (2010).

His academic affiliations include the Reinforcement Learning & Artificial Intelligence Lab (RLAI) at the University of Alberta and editorial roles at journals like Mathematics of Operations Research and the Journal of Machine Learning Research. Awards include the ECML/PKDD Test of Time Award (2016) and recognition for contributions to ICML, UAI, and other conferences. His work bridges theoretical insights with practical applications in AI, robotics, and automated decision systems.

Research interests span efficient RL algorithms, statistical guarantees for offline learning, and the intersection of optimization and uncertainty quantification. Collaborations with institutions like DeepMind emphasize foundational advances in AI, while his teaching spans graduate and undergraduate courses in machine learning and algorithms.

Key achievements include pioneering contributions to exploration strategies in bandits and RL, developing PAC-Bayes bounds for neural networks, and advancing understanding of double descent phenomena in overparameterized models. His lab (RLAI) is a hub for cutting-edge research in RL theory and applications.