Fetching the paper…

Variance-Aware Regret Bounds for Undiscounted Reinforcement Learning in MDPs · Around