2014 (modified: 11 Nov 2022)ICML 2014Readers: Everyone
Abstract:We consider stochastic multi-armed bandit problems with complex actions over a set of basic arms, where the decision maker plays a complex action rather than a basic arm in each round. The reward o...