A Near-optimal Algorithm for Learning Margin Halfspaces with Massart Noise

Abstract

We study the problem of PAC learning ${\hat{I}}^3$-margin halfspaces in the presence of Massart noise. Without computational considerations, the sample complexity of this learning problem is known to be $\widetildeΘ(1/(γ^2 ε))$. Prior computationally efficient algorithms for the problem incur sample complexity $\tilde{O}(1/(γ^4 ε^3))$ and achieve 0-1 error of $\hat{I}\cdot +\hat{I}\mu$, where $\hat{I}\cdot <1/2$ is the upper bound on the noise rate. Recent work gave evidence of an information-computation tradeoff, suggesting that a quadratic dependence on $1/\hat{I}\mu$ is required for computationally efficient algorithms. Our main result is a computationally efficient learner with sample complexity $\widetildeΘ(1/(γ^2 ε^2))$, nearly matching this lower bound. In addition, our algorithm is simple and practical, relying on online SGD on a carefully selected sequence of convex losses.