11
13

Learning Narrow One-Hidden-Layer ReLU Networks

Abstract

We consider the well-studied problem of learning a linear combination of kk ReLU activations with respect to a Gaussian distribution on inputs in dd dimensions. We give the first polynomial-time algorithm that succeeds whenever kk is a constant. All prior polynomial-time learners require additional assumptions on the network, such as positive combining coefficients or the matrix of hidden weight vectors being well-conditioned. Our approach is based on analyzing random contractions of higher-order moment tensors. We use a multi-scale analysis to argue that sufficiently close neurons can be collapsed together, sidestepping the conditioning issues present in prior work. This allows us to design an iterative procedure to discover individual neurons.

View on arXiv
Comments on this paper