ReLU Regression with Massart Noise

Abstract

We study the fundamental problem of ReLU regression, where the goal is to fit Rectified Linear Units (ReLUs) to data. This supervised learning task is efficiently solvable in the realizable setting, but is known to be computationally hard with adversarial label noise. In this work, we focus on ReLU regression in the Massart noise model, a natural and well-studied semi-random noise model. In this model, the label of every point is generated according to a function in the class, but an adversary is allowed to change this value arbitrarily with some probability, which is {\em at most} $\eta <1/2$. We develop an efficient algorithm that achieves exact parameter recovery in this model under mild anti-concentration assumptions on the underlying distribution. Such assumptions are necessary for exact recovery to be information-theoretically possible. We demonstrate that our algorithm significantly outperforms naive applications of ${\ell }_1$ and ${\ell }_2$ regression on both synthetic and real data.

References in corpus (12)

• Poisoning Attacks against Support Vector Machines
• Robust Regression via Hard Thresholding
• Reliably Learning the ReLU in Polynomial Time
• Efficient active learning of sparse halfspaces with arbitrary bounded noise
• Adaptive Hard Thresholding for Near-optimal Consistent Robust Regression
• Time/Accuracy Tradeoffs for Learning a ReLU with respect to Gaussian Marginals
• Online Robust Regression via SGD on the l1 loss
• Approximation Schemes for ReLU Regression
• On Radial Isotropic Position: Theory and Algorithms
• Efficient Learning of Linear Separators under Bounded Noise
• The Optimality of Polynomial Regression for Agnostic Learning under Gaussian Marginals
• Forster Decomposition and Learning Halfspaces with Noise