Letter on Convergence of In-Parameter-Linear Nonlinear Neural Architectures With Gradient Learnings

Ivo Bukovsky; Gejza Dohnal; Peter M. Benes; Kei Ichiji; Noriyasu Homma

doi:10.1109/TNNLS.2021.3123533

Letter on Convergence of In-Parameter-Linear Nonlinear Neural Architectures With Gradient Learnings

Ivo Bukovsky, Gejza Dohnal, Peter M. Benes, Kei Ichiji, Noriyasu Homma

MED - Health Sciences

Research output: Contribution to journal › Article › peer-review

Abstract

This letter summarizes and proves the concept of bounded-input bounded-state (BIBS) stability for weight convergence of a broad family of in-parameter-linear nonlinear neural architectures (IPLNAs) as it generally applies to a broad family of incremental gradient learning algorithms. A practical BIBS convergence condition results from the derived proofs for every individual learning point or batches for real-time applications.

Original language	English
Journal	IEEE Transactions on Neural Networks and Learning Systems
DOIs	https://doi.org/10.1109/TNNLS.2021.3123533
Publication status	Accepted/In press - 2021

Keywords

Bounded-input bounded-state stability (BIBS)
Convergence
extreme learning machines
in-parameter-linear nonlinear neural architectures (IPLNAs)
incremental gradient learnings
input-to-state stability (ISS)
Kernel
Learning systems
Loss measurement
Mechanical engineering
Neural networks
polynomial neural networks
random vector functional link networks
Stability criteria
weight convergence.

ASJC Scopus subject areas

Software
Computer Science Applications
Computer Networks and Communications
Artificial Intelligence

Access to Document

10.1109/TNNLS.2021.3123533

Cite this

@article{7eaf588913434127b10ce10c3bbb6a9a,

title = "Letter on Convergence of In-Parameter-Linear Nonlinear Neural Architectures With Gradient Learnings",

abstract = "This letter summarizes and proves the concept of bounded-input bounded-state (BIBS) stability for weight convergence of a broad family of in-parameter-linear nonlinear neural architectures (IPLNAs) as it generally applies to a broad family of incremental gradient learning algorithms. A practical BIBS convergence condition results from the derived proofs for every individual learning point or batches for real-time applications.",

keywords = "Bounded-input bounded-state stability (BIBS), Convergence, extreme learning machines, in-parameter-linear nonlinear neural architectures (IPLNAs), incremental gradient learnings, input-to-state stability (ISS), Kernel, Learning systems, Loss measurement, Mechanical engineering, Neural networks, polynomial neural networks, random vector functional link networks, Stability criteria, weight convergence.",

author = "Ivo Bukovsky and Gejza Dohnal and Benes, {Peter M.} and Kei Ichiji and Noriyasu Homma",

note = "Publisher Copyright: IEEE",

year = "2021",

doi = "10.1109/TNNLS.2021.3123533",

language = "English",

journal = "IEEE Transactions on Neural Networks and Learning Systems",

issn = "2162-237X",

publisher = "IEEE Computational Intelligence Society",

}

TY - JOUR

T1 - Letter on Convergence of In-Parameter-Linear Nonlinear Neural Architectures With Gradient Learnings

AU - Bukovsky, Ivo

AU - Dohnal, Gejza

AU - Benes, Peter M.

AU - Ichiji, Kei

AU - Homma, Noriyasu

N1 - Publisher Copyright: IEEE

PY - 2021

Y1 - 2021

N2 - This letter summarizes and proves the concept of bounded-input bounded-state (BIBS) stability for weight convergence of a broad family of in-parameter-linear nonlinear neural architectures (IPLNAs) as it generally applies to a broad family of incremental gradient learning algorithms. A practical BIBS convergence condition results from the derived proofs for every individual learning point or batches for real-time applications.

AB - This letter summarizes and proves the concept of bounded-input bounded-state (BIBS) stability for weight convergence of a broad family of in-parameter-linear nonlinear neural architectures (IPLNAs) as it generally applies to a broad family of incremental gradient learning algorithms. A practical BIBS convergence condition results from the derived proofs for every individual learning point or batches for real-time applications.

KW - Bounded-input bounded-state stability (BIBS)

KW - Convergence

KW - extreme learning machines

KW - in-parameter-linear nonlinear neural architectures (IPLNAs)

KW - incremental gradient learnings

KW - input-to-state stability (ISS)

KW - Kernel

KW - Learning systems

KW - Loss measurement

KW - Mechanical engineering

KW - Neural networks

KW - polynomial neural networks

KW - random vector functional link networks

KW - Stability criteria

KW - weight convergence.

UR - http://www.scopus.com/inward/record.url?scp=85121379463&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85121379463&partnerID=8YFLogxK

U2 - 10.1109/TNNLS.2021.3123533

DO - 10.1109/TNNLS.2021.3123533

M3 - Article

AN - SCOPUS:85121379463

JO - IEEE Transactions on Neural Networks and Learning Systems

JF - IEEE Transactions on Neural Networks and Learning Systems

SN - 2162-237X

ER -

Letter on Convergence of In-Parameter-Linear Nonlinear Neural Architectures With Gradient Learnings

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this