Publication detail

Mapping trained neural networks to FPNNs

KRČMA, M. KAŠTIL, J. KOTÁSEK, Z.

Original Title

Mapping trained neural networks to FPNNs

Type

conference paper

Language

English

Original 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.

Keywords

FPNA, FPNN, Artificial neural networks, FPGA

Authors

KRČMA, M.; KAŠTIL, J.; KOTÁSEK, Z.

RIV year

2015

Released

22. 4. 2015

Publisher

IEEE Computer Society

Location

Belgrade

ISBN

978-1-4799-6779-7

Book

IEEE 18th International Symposium on Design and Diagnostics of Electronic Circuits and Systems

Pages from

157

Pages to

160

Pages count

4

BibTex

@inproceedings{BUT119876,
  author="Martin {Krčma} and Jan {Kaštil} and Zdeněk {Kotásek}",
  title="Mapping trained neural networks to FPNNs",
  booktitle="IEEE 18th International Symposium on Design and Diagnostics of Electronic Circuits and Systems",
  year="2015",
  pages="157--160",
  publisher="IEEE Computer Society",
  address="Belgrade",
  doi="10.1109/DDECS.2015.50",
  isbn="978-1-4799-6779-7"
}