A Differential Topological View of Challenges in Learning with Feedforward Neural Networks

Abstract

Among many unsolved puzzles in theories of Deep Neural Networks (DNNs), there are three most fundamental challenges that highly demand solutions, namely, expressibility, optimisability, and generalisability. Although there have been significant progresses in seeking answers using various theories, e.g. information bottleneck theory, sparse representation, statistical inference, Riemannian geometry, etc., so far there is no single theory that is able to provide solutions to all these challenges. In this work, we propose to engage the theory of differential topology to address the three problems. By modelling the dataset of interest as a smooth manifold, DNNs can be considered as compositions of smooth maps between smooth manifolds. Specifically, our work offers a differential topological view of loss landscape of DNNs, interplay between width and depth in expressibility, and regularisations for generalisability. Finally, in the setting of deep representation learning, we further apply the quotient topology to investigate the architecture of DNNs, which enables to capture nuisance factors in data with respect to a specific learning task.

17 pages, 3 figures

References in corpus (9)

• Robust Large Margin Deep Neural Networks
• The loss surface of deep and wide neural networks
• Path-SGD: Path-Normalized Optimization in Deep Neural Networks
• The Expressive Power of Neural Networks: A View from the Width
• On the Optimization of Deep Networks: Implicit Acceleration by Overparameterization
• The power of deeper networks for expressing natural functions
• To understand deep learning we need to understand kernel learning
• i-RevNet: Deep Invertible Networks
• Neural Networks Should Be Wide Enough to Learn Disconnected Decision Regions

Cited by in corpus (1)

• Analysis and Optimisation of Bellman Residual Errors with Neural Function Approximation