Article
Computational aspects of Kolmogorov's superposition theorem
Neural Networks
(1994)
Abstract
This paper continues the investigation of representations of continuous functions f(x1, …, xn) with n ≥ 2 in the form f(X1,…,Xn) = ∑2nq=0Φq[∑np=1λ1ψ(xp + qen with a predetermined function ψ that is independent of n. The function ψ is defined through its graph that is the limit point of iterated contraction mappings. The functions ψ and Φq are the uniform limits of sequences of computable functions constructed with a fixed mapping σ, which itself can be approximated with sigmoid functions.
Keywords
- Superpositions,
- Kolmogorov,
- Contraction mapping,
- Representation of continuous functions of several variables,
- Approximation of functions
Disciplines
Publication Date
1994
Publisher Statement
SJSU users: use the following link to login and access the article via SJSU databases
Citation Information
Hidefumi Katsuura and David A. Sprecher. "Computational aspects of Kolmogorov's superposition theorem" Neural Networks Vol. 7 Iss. 3 (1994) Available at: http://works.bepress.com/hidefumi_katsuura/7/