[2403.06806] On the Global Convergence of Policy Gradient in Average Reward Markov Decision Processes