Ergodicity in reinforcement learning
Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences · 19.11.2026 · 384 (2334)
DOI: 10.1098/rsta.2025.0156 ↗Kurz gefasst
Fachartikel in der Zeitschrift Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, online erschienen am 19.11.2026. Verfasst von Dominik Baumann und 7 weiteren Autor:innen. Die Arbeit nennt 58 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Finnish Ministry of Education and Culture, Research Council of Finland Flagship programme: Finnish Center for Artificial Intelligence FCAI, Kjell och Märta Beijer Foundation.
Abstract (Originalsprache)
In reinforcement learning, we typically aim to optimize the expected value of the sum of rewards an agent collects over a trajectory. However, if the process generating these rewards is non-ergodic, the expected value, i.e. the average over infinitely many trajectories with a given policy, is uninformative for the average over a single, but infinitely long trajectory. Thus, if we care about how the individual agent performs during deployment, the expected value is not a good optimization objective. In this paper, we discuss the impact of non-ergodic reward processes on reinforcement learning agents through an instructive example, relate the notion of ergodic reward processes to more widely used notions of ergodic Markov chains, and present existing solutions that optimize long-term performance of individual trajectories under non-ergodic reward dynamics. This article is part of the theme issue ‘Ergodicity and ergodicity breaking in the sciences’.
Förderung
- Finnish Ministry of Education and Culture
- Research Council of Finland Flagship programme: Finnish Center for Artificial Intelligence FCAI
- Kjell och Märta Beijer Foundation
Zitieren
APA
Baumann, D., Noorani, E., Mustafin, A., Sheng, X., Verbruggen, B., Vanhoyweghen, A., Ginis, V., & Schön, T. B. (2026). Ergodicity in reinforcement learning. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 384(2334). https://doi.org/10.1098/rsta.2025.0156
Vancouver
Baumann D, Noorani E, Mustafin A, Sheng X, Verbruggen B, Vanhoyweghen A, et al. Ergodicity in reinforcement learning. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences. 2026;384(2334). doi:10.1098/rsta.2025.0156
BibTeX
@article{Baumann2026,
author = {Baumann, Dominik and Noorani, Erfaun and Mustafin, Arsenii and Sheng, Xinyi and Verbruggen, Bert and Vanhoyweghen, Arne and Ginis, Vincent and Schön, Thomas B.},
title = {Ergodicity in reinforcement learning},
journal = {Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences},
year = {2026},
volume = {384},
number = {2334},
doi = {10.1098/rsta.2025.0156}
}
EUVIMED