Networked Bandits with Disjoint Linear Payoffs
Published on Oct 07, 20141933 Views
In this paper, we study `networked bandits', a new bandit problem where a set of interrelated arms varies over time and, given the contextual information that selects one arm, invokes other correlated