SCALABLE ACTIVATION FUNCTION EMPLOYMENT IN PHYSICS INFORMED HIGHER ORDER NEURAL NETWORKS FOR INITIAL VALUE PROBLEM APPROXIMATION

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21220%2F23%3A00370807" target="_blank" >RIV/68407700:21220/23:00370807 - isvavai.cz</a>

Výsledek na webu

<a href="https://isimconference.eu/" target="_blank" >https://isimconference.eu/</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

SCALABLE ACTIVATION FUNCTION EMPLOYMENT IN PHYSICS INFORMED HIGHER ORDER NEURAL NETWORKS FOR INITIAL VALUE PROBLEM APPROXIMATION

Popis výsledku v původním jazyce

There are many physical phenomena in the engineering applications, which are governed by ordinary differential equations. In the practice these problems are approximated using classical numerical methods. Present contribution deals with an alternative approximation method – higher order neural networks and furthermore, employment of the scalable activation function which can accelerate speed of the convergence is discussed. Basic ideas of the methods are introduced and designed architectures are tested on initial value problem. A comparison of the results quality between three neural network architectures is performed. Improvement in the speed of the convergence has been reached.

Název v anglickém jazyce

SCALABLE ACTIVATION FUNCTION EMPLOYMENT IN PHYSICS INFORMED HIGHER ORDER NEURAL NETWORKS FOR INITIAL VALUE PROBLEM APPROXIMATION

Popis výsledku anglicky

There are many physical phenomena in the engineering applications, which are governed by ordinary differential equations. In the practice these problems are approximated using classical numerical methods. Present contribution deals with an alternative approximation method – higher order neural networks and furthermore, employment of the scalable activation function which can accelerate speed of the convergence is discussed. Basic ideas of the methods are introduced and designed architectures are tested on initial value problem. A comparison of the results quality between three neural network architectures is performed. Improvement in the speed of the convergence has been reached.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/EF16_019%2F0000826" target="_blank" >EF16_019/0000826: Centrum pokročilých leteckých technologií</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)<br>S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

BOOK OF ABSTRACTS ISIM&ISWIM

ISBN

—

ISSN

2821-8779

e-ISSN

2821-8779

Počet stran výsledku

2

Strana od-do

88-89

Název nakladatele

1 Decembrie 1918 University

Místo vydání

Alba Iulia

Místo konání akce

Bucharest

Datum konání akce

28. 6. 2023

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—