The Power of Localization for Efficiently Learning Linear Separators with Noise

TL;DR

This paper presents new noise-tolerant algorithms for efficiently learning linear separators in high-dimensional spaces, achieving near-optimal noise robustness and polylogarithmic label complexity in active learning scenarios.

Contribution

It introduces the first polynomial-time algorithms for learning linear separators under malicious and adversarial label noise with improved noise tolerance guarantees.

Findings

Algorithms tolerate nearly optimal noise rates of () in both noise models.

Achieves polylogarithmic label complexity in active learning for noisy linear separators.

Provides the first polynomial-time active learning algorithms under malicious and adversarial noise.

Abstract

We introduce a new approach for designing computationally efficient learning algorithms that are tolerant to noise, and demonstrate its effectiveness by designing algorithms with improved noise tolerance guarantees for learning linear separators. We consider both the malicious noise model and the adversarial label noise model. For malicious noise, where the adversary can corrupt both the label and the features, we provide a polynomial-time algorithm for learning linear separators in $\Re^d$ under isotropic log-concave distributions that can tolerate a nearly information-theoretically optimal noise rate of $\eta = \Omega(\epsilon)$. For the adversarial label noise model, where the distribution over the feature vectors is unchanged, and the overall probability of a noisy label is constrained to be at most $\eta$, we also give a polynomial-time algorithm for learning linear separators in $\Re^d$ under isotropic log-concave distributions that can handle a noise rate of $\eta = \Omega\left(\epsilon\right)$. We show that, in the active learning model, our algorithms achieve a label complexity whose dependence on the error parameter $\epsilon$ is polylogarithmic. This provides the first polynomial-time active learning algorithm for learning linear separators in the presence of malicious noise or adversarial label noise.