Regret bounds for Narendra-Shapiro bandit algorithms | Read Paper on Bytez