On the Sample Complexity of Two-Layer Networks: Lipschitz vs. Element-Wise Lipschitz Activation

TL;DR

This paper analyzes the sample complexity of two-layer neural networks with Lipschitz activation functions, showing that element-wise activations lead to logarithmic width dependency, unlike non-element-wise functions.

Contribution

It establishes the importance of element-wise activation functions for achieving logarithmic sample complexity dependency on network width.

Findings

Logarithmic sample complexity for element-wise activations

Linear width dependency for certain non-element-wise activations

Development of new techniques using Approximate Description Length (ADL)

Abstract

We investigate the sample complexity of bounded two-layer neural networks using different activation functions. In particular, we consider the class $$ \mathcal{H} = \left\{\textbf{x}\mapsto \langle \textbf{v}, \sigma \circ W\textbf{b} + \textbf{b} \rangle : \textbf{b}\in\mathbb{R}^d, W \in \mathbb{R}^{\mathcal{T}\times d}, \textbf{v} \in \mathbb{R}^{\mathcal{T}}\right\} $$ where the spectral norm of $W$ and $\textbf{v}$ is bounded by $O(1)$, the Frobenius norm of $W$ is bounded from its initialization by $R > 0$, and $\sigma$ is a Lipschitz activation function. We prove that if $\sigma$ is element-wise, then the sample complexity of $\mathcal{H}$ has only logarithmic dependency in width and that this complexity is tight, up to logarithmic factors. We further show that the element-wise property of $\sigma$ is essential for a logarithmic dependency bound in width, in the sense that there exist non-element-wise activation functions whose sample complexity is linear in width, for widths that can be up to exponential in the input dimension. For the upper bound, we use the recent approach for norm-based bounds named Approximate Description Length (ADL) by arXiv:1910.05697. We further develop new techniques and tools for this approach that will hopefully inspire future works.