Piecewise linear functions and neural network expressivity via discriminantal arrangements

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autor principal: Das, Pragnya
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911564318113792
author Das, Pragnya
author_facet Das, Pragnya
contents We extend the hyperplane arrangement framework for neural network expressivity from the braid to discriminantal arrangements. Compatible piecewise linear functions are characterized by circuit relations and admit a matroidal description via Mobius inversion, with dimension equal to the number of independent sets. For circuits of size three, functions are determined by values on subsets of size at most two.
format Preprint
id arxiv_https___arxiv_org_abs_2604_02480
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Piecewise linear functions and neural network expressivity via discriminantal arrangements
Das, Pragnya
Combinatorics
Algebraic Geometry
05B35, 52C35(Primary), 68T07 (Secondary)
We extend the hyperplane arrangement framework for neural network expressivity from the braid to discriminantal arrangements. Compatible piecewise linear functions are characterized by circuit relations and admit a matroidal description via Mobius inversion, with dimension equal to the number of independent sets. For circuits of size three, functions are determined by values on subsets of size at most two.
title Piecewise linear functions and neural network expressivity via discriminantal arrangements
topic Combinatorics
Algebraic Geometry
05B35, 52C35(Primary), 68T07 (Secondary)
url https://arxiv.org/abs/2604.02480