Bandits, Query Learning, and the Haystack Dimension

Kareem Amin, Michael Kearns, Umar Syed ; JMLR W&CP 19:87-106, 2011.

Abstract

Motivated by multi-armed bandits (MAB) problems with a very large or even infinite number of arms, we consider the problem of finding a maximum of an unknown target function by querying the function at chosen inputs (or arms).We give an analysis of the query complexity of this problem, under the assumption that the payoff of each arm is given by a function belonging to a known, finite, but otherwise arbitrary function class. Our analysis centers on a new notion of function class complexity that we callthe \emph{haystack dimension}, which is used to prove the approximate optimality of a simple greedy algorithm. This algorithm is then usedas a subroutine in a \parametric MAB algorithm, yielding provably near-optimal regret. We provide a generalization to the infinite cardinality setting, andcomment on how our analysis is connected to, and improves upon, existing results for query learning and generalized binary search.

Page last modified on Sat Dec 17 01:04 CET 2011.