Paper COLT 2025 Bandits and online learning

Logarithmic Regret for Convex–Concave Saddle Problems

Maya Singh, Renat Ostrovsky, Leila Aydın · OOAARG · Department of Computer Science

When both players play strongly convex losses, an extragradient method attains O(logT)O(\log T) duality gap with high probability.