Skip to yearly menu bar Skip to main content


Poster

Learning One-hidden-layer Neural Networks with Landscape Design

Rong Ge · Jason Lee · Tengyu Ma

East Meeting level; 1,2,3 #13

Abstract: We consider the problem of learning a one-hidden-layer neural network: we assume the input x is from Gaussian distribution and the label y=aσ(Bx)+ξ, where a is a nonnegative vector and B is a full-rank weight matrix, and ξ is a noise vector. We first give an analytic formula for the population risk of the standard squared loss and demonstrate that it implicitly attempts to decompose a sequence of low-rank tensors simultaneously. Inspired by the formula, we design a non-convex objective function G whose landscape is guaranteed to have the following properties: 1. All local minima of G are also global minima. 2. All global minima of G correspond to the ground truth parameters. 3. The value and gradient of G can be estimated using samples. With these properties, stochastic gradient descent on G provably converges to the global minimum and learn the ground-truth parameters. We also prove finite sample complexity results and validate the results by simulations.

Live content is unavailable. Log in and register to view live content