A. Klivans and P. Long and R. Servedio.

Preliminary version in

We give new algorithms for learning halfspaces in the challenging {\it malicious noise} model, where an adversary may corrupt both the labels and the underlying distribution of examples. Our algorithms can tolerate malicious noise rates exponentially larger than previous work in terms of the dependence on the dimension $n$, and succeed for the fairly broad class of all isotropic log-concave distributions.

We give poly$(n, 1/\eps)$-time algorithms for solving the following problems to accuracy $\epsilon$:

We also give a poly$(n,1/\eps)$-time algorithm for learning origin-centered halfspaces under any isotropic log-concave distribution on $\R^n$ in the presence of \emph{adversarial label noise} at rate $\eta = \Omega(\eps^{2}/\log(1/\eps))$. In the adversarial label noise setting (or agnostic model), labels can be noisy, but not example points themselves. Previous results could handle $\eta = \Omega(\eps)$ but had running time exponential in an unspecified function of $1/\eps$.

Our analysis crucially exploits both concentration and anti-concentration properties of isotropic log-concave distributions. Our algorithms combine an iterative outlier removal procedure using Principal Component Analysis together with ``smooth'' boosting.

pdf of conference version