English

Truth or Backpropaganda? An Empirical Investigation of Deep Learning Theory

Machine Learning 2020-04-29 v3 Optimization and Control Machine Learning

Abstract

We empirically evaluate common assumptions about neural networks that are widely held by practitioners and theorists alike. In this work, we: (1) prove the widespread existence of suboptimal local minima in the loss landscape of neural networks, and we use our theory to find examples; (2) show that small-norm parameters are not optimal for generalization; (3) demonstrate that ResNets do not conform to wide-network theories, such as the neural tangent kernel, and that the interaction between skip connections and batch normalization plays a role; (4) find that rank does not correlate with generalization or robustness in a practical setting.

Keywords

Cite

@article{arxiv.1910.00359,
  title  = {Truth or Backpropaganda? An Empirical Investigation of Deep Learning Theory},
  author = {Micah Goldblum and Jonas Geiping and Avi Schwarzschild and Michael Moeller and Tom Goldstein},
  journal= {arXiv preprint arXiv:1910.00359},
  year   = {2020}
}

Comments

18 pages, 6 figures. First two authors contributed equally. Published as a conference paper at ICLR 2020