HOLDER CONTINUOUS ACTIVATION FUNCTIONS IN NEURAL NETWORKS

Nasser-Eddine Mohamed Ali Tatar

HOLDER CONTINUOUS ACTIVATION FUNCTIONS IN NEURAL NETWORKS

Nasser-Eddine Mohamed Ali Tatar

Department of Mathematics

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

Abstract

We consider a system which arises in Neural Network Theory and prove an exponential stabilization result in case of Holder continuous activation functions. This extends the previous works where activation functions are assumed to be at least Lipschitz continuous.

Original language	English
Journal	PUSHPA PUBLISHING HOUSE
State	Published - 2015

Cite this

@article{87e4cb2d5a124fec8eef0ae3ee0fa18f,

title = "HOLDER CONTINUOUS ACTIVATION FUNCTIONS IN NEURAL NETWORKS",

abstract = "We consider a system which arises in Neural Network Theory and prove an exponential stabilization result in case of Holder continuous activation functions. This extends the previous works where activation functions are assumed to be at least Lipschitz continuous.",

author = "Tatar, \{Nasser-Eddine Mohamed Ali\}",

year = "2015",

language = "English",

}

TY - JOUR

T1 - HOLDER CONTINUOUS ACTIVATION FUNCTIONS IN NEURAL NETWORKS

AU - Tatar, Nasser-Eddine Mohamed Ali

PY - 2015

Y1 - 2015

N2 - We consider a system which arises in Neural Network Theory and prove an exponential stabilization result in case of Holder continuous activation functions. This extends the previous works where activation functions are assumed to be at least Lipschitz continuous.

AB - We consider a system which arises in Neural Network Theory and prove an exponential stabilization result in case of Holder continuous activation functions. This extends the previous works where activation functions are assumed to be at least Lipschitz continuous.

M3 - Article

SN - 0974-3243

JO - PUSHPA PUBLISHING HOUSE

JF - PUSHPA PUBLISHING HOUSE

ER -

HOLDER CONTINUOUS ACTIVATION FUNCTIONS IN NEURAL NETWORKS

Abstract

Fingerprint

Cite this