Provably Efficient RL under Episode-Wise Safety in Constrained MDPs with Linear Function Approximation

NeurIPSSpotlight2025

Authors: Toshinori Kitamura, Arnob Ghosh, Tadashi Kozuno, Wataru Kumagai, Kazumi Kasaura, Kenta Hoshino, Yohei Hosoe, Yutaka Matsuo
Venue: NeurIPS 2025
Track: Spotlight

TL;DR

We study the reinforcement learning (RL) problem in a constrained Markov decision process (CMDP), where an agent explores the environment to maximize the expected cumulative reward while satisfying a…

Opening excerpt from the authors’ abstract. source

Read the paper

OpenReview

Topics

reinforcement learning efficient safety agent

← All NeurIPS 2025 Spotlight papers · Browse the whole archive