Maxout Polytopes
Combinatorics
2025-09-26 v1 Discrete Mathematics
Machine Learning
Abstract
Maxout polytopes are defined by feedforward neural networks with maxout activation function and non-negative weights after the first layer. We characterize the parameter spaces and extremal f-vectors of maxout polytopes for shallow networks, and we study the separating hypersurfaces which arise when a layer is added to the network. We also show that maxout polytopes are cubical for generic networks without bottlenecks.
Keywords
Cite
@article{arxiv.2509.21286,
title = {Maxout Polytopes},
author = {Andrei Balakin and Shelby Cox and Georg Loho and Bernd Sturmfels},
journal= {arXiv preprint arXiv:2509.21286},
year = {2025}
}
Comments
24 pages, 3 figures