Csaba Szepesvári
Also known as Csaba Szepesvari, Csaba Szepesvàri
Csaba Szepesvári develops algorithms for learning and decision-making. His work includes shared invention of UCT and books on reinforcement learning and bandits. Szepesvári completed his doctorate in probability and statistics at József Attila University in Szeged, Hungary, in 1999. His later research focused on principled learning methods for artificial intelligence, especially algorithms for sequential decisions. Szepesvári co-founded the Reinforcement Learning Theory virtual seminar series in 2020. He served as an ICML program co-chair in 2022. These activities complemented his work on the foundations of learning algorithms.
- Now at
- University of Alberta
- Works
- 438
- Citations
- 16,575
- h-index
- 52
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
1999
Training in probability and statistics
Szepesvári completed his doctorate in probability and statistics at József Attila University in Szeged, Hungary, in 1999. His later research focused on principled learning methods for artificial intelligence, especially algorithms for sequential decisions.
-
2006
A search method for uncertain decisions
Szepesvári co-invented UCT, a Monte Carlo tree-search algorithm. AlphaGo used a variant ten years later, in its 2016 match with Lee Sedol.
-
2010–2020
Explaining reinforcement learning
He wrote Algorithms for Reinforcement Learning in 2010. In 2020, he and Tor Lattimore published Bandit Algorithms. The books describe two related areas of learning-based decision-making.
-
2020–2022
Building a research community
Szepesvári co-founded the Reinforcement Learning Theory virtual seminar series in 2020. He served as an ICML program co-chair in 2022. These activities complemented his work on the foundations of learning algorithms.
Sources and credits
Photo: ellis.eu; Csaba Szepesvári profile photograph; creator not stated. License: Not stated by source. Source: ellis.eu. Cropped to a square and resized. Ask us to remove or replace this photo.