Abstract
In order to ride on the strength of parallel operation which is feature of Neural Network (NN), It should be implemented on hardware. Considering hardware implementation of NN, it is preferable that the circuit scale of each neuron is compact. We have been proposed Pulsed NN based on Delta-Sigma Modulation (DSM-PNN). The Delta-Sigma Modulator (DSM), which is a key of our proposed DSM-PNN, is well known as a method to convert the input signal into 1-bit pulse stream. DSM-PNN copes with both operating accuracy and compact circuit scale, Because PNN has an advantage of whose circuit scale is compact, and a accuracy of operation. In this paper, an outline of DSM-PNN and how to realize a learning rule of NN are explained by taking ICA as example.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Murahashi, Y., Hotta, H., Doki, S., Okuma, S.: Hardware Realization of Novel Pulsed Neural Networks Based on Delta-Sigma Modulation with GHA Learning Rule. In: The 2002 IEEE Asia-Pacific Conference on Circuits and Systems,
Murahashi, Y., Hotta, H., Doki, S., Okuma, S.: Pulsed Neural Networks Based on Delta-Sigma Modulation with GHA Learning Rule and Their Hardware Implementation. The IEICE Transactions on Information and Systems, PT.2 J87-D-2(2), 705–715 (2004) (Japanese)
Amari, S.: Natural gradien learning works efficiently in learning Neural Computation (1997)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Hotta, H., Murahashi, Y., Doki, S., Okuma, S. (2004). Hardware Implementation of Pulsed Neural Networks Based on Delta-Sigma Modulator for ICA. In: Yin, FL., Wang, J., Guo, C. (eds) Advances in Neural Networks – ISNN 2004. ISNN 2004. Lecture Notes in Computer Science, vol 3173. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-28647-9_117
Download citation
DOI: https://doi.org/10.1007/978-3-540-28647-9_117
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-22841-7
Online ISBN: 978-3-540-28647-9
eBook Packages: Springer Book Archive