We show that if the closureof a function class F under the metric induced by some probability distribution is not convex, then the sample complexity for agnostically learning F wi...
Wee Sun Lee, Peter L. Bartlett, Robert C. Williams...
We present and analyze an agnostic active learning algorithm that works without keeping a version space. This is unlike all previous approaches where a restricted set of candidate...
Alina Beygelzimer, Daniel Hsu, John Langford, Tong...
We prove that the concept class of disjunctions cannot be pointwise approximated by linear combinations of any small set of arbitrary real-valued functions. That is, suppose that t...
In this paper we initiatean investigationof generalizationsof the ProbablyApproximatelyCorrect (PAC) learningmodelthat attemptto significantlyweakenthe target functionassumptions.T...
Michael J. Kearns, Robert E. Schapire, Linda Selli...
Abstract— The optimal model parameters of a kernel machine are typically given by the solution of a convex optimisation problem with a single global optimum. Obtaining the best p...