Paperlane

no. 3 · Tuesday 27 October 2026

On shallow feedforward neural networks with inputs from a topological space

Vugar Ismailov

math.FA · v3

Abstract

We study feedforward neural networks with inputs from a topological space (TFNNs). We prove a universal approximation theorem for shallow TFNNs, which demonstrates their capacity to approximate any continuous function defined on this topological space. As an application, we obtain an approximative version of Kolmogorov's superposition theorem for compact metric spaces.

comments
Published version with two minor corrections. Namely, in Theorem 2.1, C(K) was replaced with C(X), and in the proof of Proposition 2.1, the formula for τ(t) was corrected. The results remain unchanged
journal
V.E. Ismailov, Ann. Math. Artif. Intell. (2026)
fields
Functional Analysis
pdf ↗ page arXiv:2504.02321 ↗