Evolution Strategies for Deep Neural Network Models Design
The result's identifiers
Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F67985807%3A_____%2F17%3A00478624" target="_blank" >RIV/67985807:_____/17:00478624 - isvavai.cz</a>
Result on the web
<a href="http://ceur-ws.org/Vol-1885/159.pdf" target="_blank" >http://ceur-ws.org/Vol-1885/159.pdf</a>
DOI - Digital Object Identifier
—
Alternative languages
Result language
angličtina
Original language name
Evolution Strategies for Deep Neural Network Models Design
Original language description
Deep neural networks have become the state-of art methods in many fields of machine learning recently. Still, there is no easy way how to choose a network architecture which can significantly influence the network performance. This work is a step towards an automatic architecture design. We propose an algorithm for an optimization of a network architecture based on evolution strategies. The algorithm is inspired by and designed directly for the Keras library [3] which is one of the most common implementations of deep neural networks. The proposed algorithm is tested on MNIST data set and the prediction of air pollution based on sensor measurements, and it is compared to several fixed architectures and support vector regression.
Czech name
—
Czech description
—
Classification
Type
D - Article in proceedings
CEP classification
—
OECD FORD branch
10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)
Result continuities
Project
<a href="/en/project/GA15-18108S" target="_blank" >GA15-18108S: Model complexity of neural, radial, and kernel networks</a><br>
Continuities
P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)
Others
Publication year
2017
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Data specific for result type
Article name in the collection
Proceedings ITAT 2017: Information Technologies - Applications and Theory
ISBN
978-1974274741
ISSN
1613-0073
e-ISSN
—
Number of pages
8
Pages from-to
159-166
Publisher name
Technical University & CreateSpace Independent Publishing Platform
Place of publication
Aachen & Charleston
Event location
Martinské hole
Event date
Sep 22, 2017
Type of event by nationality
EUR - Evropská akce
UT code for WoS article
—