Optimal High-Probability Regret for Online Convex Optimization with Two-Point Bandit Feedback | Signal Canvas

Optimal High-Probability Regret for Online Convex Optimization with Two-Point Bandit Feedback | Signal Canvas | ScienceToStartup