James S. Plank, ChaoHui Zheng, Catherine D. Schuman and Christopher Dean

July, 2021

ICONS: International Conference on Neuromorphic Systems

https://icons.ornl.gov

View Article

Abstract

In this paper, we focus on the hand construction of small-scale, spiking, neuromorphic networks. They are partitioned into two sets. The first set performs the binary operations AND, OR and XOR. For each of these operations, there are eight scenarios that we consider, with four types of encodings of binary values to spikes, and neurons that feature or prohibit leak. The second set of networks perform conversions of binary values from one type of encoding to another, again considering both leak or no leak. These networks are presented graphically and with spike-raster plots that illustrate their activity. We tabulate metrics of interest concerning the size, activity and speed of these networks. Our goal with this work is to aid in the testing of neuromorphic hardware, and to enable the composition of multiple spiking neural networks, perhaps trained with other methodologies, without requiring information to leave a neuroprocessor for processing by conventional hardware.

Citation Information

Text

author           J. S. Plank and C. Zheng and C. D. Schuman and C. Dean
title            Spiking Neuromorphic Networks for Binary Tasks
booktitle        International Conference on Neuromorphic Computing Systems (ICONS)
publisher        ACM
pages            1-8
year             2021

Bibtex

@INPROCEEDINGS{pzs:21:snn,
    author = "J. S. Plank and C. Zheng and C. D. Schuman and C. Dean",
    title = "Spiking Neuromorphic Networks for Binary Tasks",
    booktitle = "International Conference on Neuromorphic Computing Systems (ICONS)",
    publisher = "ACM",
    pages = "1-8",
    year = "2021"
}