2018 (modified: 04 Nov 2022)AISTATS 2018Readers: Everyone
Abstract:We consider a variant of the classic multi-armed bandit problem where the expected reward of each arm is a function of an unknown parameter. The arms are divided into different groups, each of whic...