Paper accepted at ICML 2026
in News on Reinforcement learning, State abstraction, Category theory, Research, Publication
Compositional behavioral semantics for state abstraction in reinforcement learning.
Manuel’s work with Yivan Zhang and Ziyan Luo was accepted at the International Conference on Machine Learning 2026. The paper develops a compositional framework for identifying behavioral structures preserved by state abstraction in reinforcement learning. The poster and conference page, conference paper, and preprint are available online.
