Local and global topological complexity measures OF ReLU neural network functions
Preprint
Abstract
We apply a generalized piecewise-linear (PL) version of Morse theory due to Grunert-Kuhnel-Rote to define and study new local and global notions of topological complexity for fully-connected feedforward ReLU neural network functions, F: R^n -> R. Along the way, we show how to construct, for each such F, a canonical polytopal complex K(F) and a deformation retract of the domain onto K(F), yielding a convenient compact model for performing calculations. We also give a construction showing that local complexity can be arbitrarily high.Keywords
All Related Versions
This publication has 5 references indexed in Scilit:
- PL Morse theory in low dimensionsAdvances in Geometry, 2023
- On Transversality of Bent Hyperplane Arrangements and the Topological Expressiveness of ReLU Neural NetworksSIAM Journal on Applied Algebra and Geometry, 2022
- On the Complexity of Neural Network Classifiers: A Comparison Between Shallow and Deep ArchitecturesIEEE Transactions on Neural Networks and Learning Systems, 2014
- Computing Persistent HomologyDiscrete & Computational Geometry, 2004
- Critical points and curvature for embedded polyhedraJournal of Differential Geometry, 1967