On Sparsity in Overparametrised Shallow ReLU Networks

Abstract

The analysis of neural network training beyond their linearization regime remains an outstanding open question, even in the simplest setup of a single hidden-layer. The limit of infinitely wide networks provides an appealing route forward through the mean-field perspective, but a key challenge is to bring learning guarantees back to the finite-neuron setting, where practical algorithms operate. Towards closing this gap, and focusing on shallow neural networks, in this work we study the ability of different regularisation strategies to capture solutions requiring only a finite amount of neurons, even on the infinitely wide regime. Specifically, we consider (i) a form of implicit regularisation obtained by injecting noise into training targets [Blanc et al.~19], and (ii) the variation-norm regularisation [Bach~17], compatible with the mean-field scaling. Under mild assumptions on the activation function (satisfied for instance with ReLUs), we establish that both schemes are minimised by functions having only a finite number of neurons, irrespective of the amount of overparametrisation. We study the consequences of such property and describe the settings where one form of regularisation is favorable over the other.

submitted to Neurips'20

References in corpus (11)

• DGM: A deep learning algorithm for solving partial differential equations
• Neural Tangent Kernel: Convergence and Generalization in Neural Networks
• A Mean Field View of the Landscape of Two-Layers Neural Networks
• The generalization error of random features regression: Precise asymptotics and double descent curve
• A Modern Take on the Bias-Variance Tradeoff in Neural Networks
• Gradient Dynamics of Shallow Univariate ReLU Networks
• Gradient Descent Quantizes ReLU Network Features
• Implicit regularization for deep neural networks driven by an Ornstein-Uhlenbeck like process
• Network size and weights size for memorization with two-layers neural networks
• A Function Space View of Bounded Norm Infinite Width ReLU Nets: The Multivariate Case
• Convex Geometry and Duality of Over-parameterized Neural Networks

Cited by in corpus (3)

• A Bregman Learning Framework for Sparse Neural Networks
• A Dynamical Central Limit Theorem for Shallow Neural Networks
• A Functional Perspective on Learning Symmetric Functions with Neural Networks