P

Initializing...

Theorem 1, p. 14 — Algorithm 3 has expected α-regret at most α²R²/(Tη) + ηF²/2 + 3Fε · Prove2Me