GD doesn't make the cut: Three ways that non-differentiability affects neural network training

This paper investigates the distinctions between gradient methods applied to non-differentiable functions (NGDMs) and classical gradient descents (GDs) designed for differentiable functions. First, we demonstrate significant differences in the convergence properties of NGDMs compared to GDs, challenging the applicability of the extensive neural network convergence literature based on $L-smoothness$ to non-smooth neural networks. Next, we demonstrate the paradoxical nature of NGDM solutions for $L_{1}$-regularized problems, showing that increasing the regularization penalty leads to an increase in the $L_{1}$ norm of …

convergence cs.cv cs.lg differences differentiable functions gds gradient literature network network training neural network paper training