Math proof: simple AI learning method reliably finds best choice over time

Original: Convergence and Regret of the Policy Gradient for Multi-Armed Bandits in Diffusion Environment

Source: arxiv.org ↗