Publication Details

Mapping trained neural networks to FPNNs

KRČMA Martin, KAŠTIL Jan and KOTÁSEK Zdeněk. Mapping trained neural networks to FPNNs. In: IEEE 18th International Symposium on Design and Diagnostics of Electronic Circuits and Systems. Belgrade: IEEE Computer Society, 2015, pp. 157-160. ISBN 978-1-4799-6779-7.
Czech title
Mapování naučených neuronových sítí na FPNN
Type
conference paper
Language
english
Authors
Krčma Martin, Ing. (DCSY FIT BUT)
Kaštil Jan, Ing. (DCSY FIT BUT)
Kotásek Zdeněk, doc. Ing., CSc. (DCSY FIT BUT)
Keywords

FPNA, FPNN, Artificial neural networks, FPGA

Abstract

This paper introduces a set of methods for mapping
the trained neural networks into the lighted grid structured
Field Programmable Neural Networks without the use of a
training data set. These methods use information obtained from
original neural networks such as a network structure, connection
weights and biases. The principles of these mapping methods are
described and the used grid FPNNs are explained. The results of
experiments are presented and summarized.

Published
2015
Pages
157-160
Proceedings
IEEE 18th International Symposium on Design and Diagnostics of Electronic Circuits and Systems
Conference
IEEE International Symposium on Design and Diagnostics of Electronic Circuits and Systems 2015, Belgrade, RS
ISBN
978-1-4799-6779-7
Publisher
IEEE Computer Society
Place
Belgrade, RS
DOI
EID Scopus
BibTeX
@INPROCEEDINGS{FITPUB10918,
   author = "Martin Kr\v{c}ma and Jan Ka\v{s}til and Zden\v{e}k Kot\'{a}sek",
   title = "Mapping trained neural networks to FPNNs",
   pages = "157--160",
   booktitle = "IEEE 18th International Symposium on Design and Diagnostics of Electronic Circuits and Systems",
   year = 2015,
   location = "Belgrade, RS",
   publisher = "IEEE Computer Society",
   ISBN = "978-1-4799-6779-7",
   doi = "10.1109/DDECS.2015.50",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/10918"
}
Back to top