Beyond Random Matrix Theory for Deep Networks
arXiv:2006.07721
Abstract
We investigate whether the Wigner semi-circle and Marcenko-Pastur distributions, often used for deep neural network theoretical analysis, match empirically observed spectral densities. We find that even allowing for outliers, the observed spectral shapes strongly deviate from such theoretical predictions. This raises major questions about the usefulness of these models in deep learning. We further show that theoretical results, such as the layered nature of critical points, are strongly dependent on the use of the exact form of these limiting spectral densities. We consider two new classes of matrix ensembles; random Wigner/Wishart ensemble products and percolated Wigner/Wishart ensembles, both of which better match observed spectra. They also give large discrete spectral peaks at the origin, providing a theoretical explanation for the observation that various optima can be connected by one dimensional of low loss values. We further show that, in the case of a random matrix product, the weight of the discrete spectral component at depends on the ratio of the dimensions of the weight matrices.
arXiv comment8 pages 5 Figures
References in corpus (11)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- The Loss Surfaces of Multilayer Networks
- Essentially No Barriers in Neural Network Energy Landscape
- The loss surface of deep and wide neural networks
- Empirical Analysis of the Hessian of Over-Parametrized Neural Networks
- Distributions of Angles in Random Packing on Spheres
- Spectrum of the Product of Independent Random Gaussian Matrices
- Eigenvalues of the Hessian in Deep Learning: Singularity and Beyond
- The Full Spectrum of Deepnet Hessians at Scale: Dynamics with SGD Training and Sample Size
- Measurements of Three-Level Hierarchical Structure in the Outliers in the Spectrum of Deepnet Hessians
- Average-case Acceleration Through Spectral Density Estimation