Shelah O. Ameli, Adam Foshie, Drew Friend, James S. Plank, Garrett S. Rose and Catherine D. Schuman

August, 2023

ICONS: International Conference on Neuromorphic Systems

https://icons.ornl.gov/schedule/

View Article

Abstract

Synaptic plasticity has long been thought to be key to learning in both biological neural networks and artificial neural networks. There are a wide array of plasticity rules in biological neural systems, but neuromorphic computing has mainly focused on spike timing dependent plasticity (STDP). In this work, we evaluate the impact of a programmable STDP mechanism on the application and algorithm performance of neuromorphic hardware on several simple classification and control tasks. We evaluate three different types of STDP, as well as no STDP, and we examine their impact on application performance and network size. We show that the best performing STDP approach depends on the application, but we also find that the inclusion of STDP tends to lead to smaller networks in terms of number of synapses.

Citation Information

Text

author       S. Ameli and A. Foshie and D. Friend and J. S. Plank and G. S. Rose and C. D. Schuman
title        Algorithm and Application Impacts of Programmable Plasticity in Spiking Neuromorphic Hardware
booktitle    International Conference on Neuromorphic Computing Systems (ICONS)
doi          10.1145/3589737.3605995
url          https://doi.org/10.1145/3589737.3605995
publisher    ACM
year         2023

Bibtex

@INPROCEEDINGS{aff:23:aai,
    author = "S. Ameli and A. Foshie and D. Friend and J. S. Plank and 
                 G. S. Rose and C. D. Schuman",
    title = "Algorithm and Application Impacts of Programmable Plasticity in 
                 Spiking Neuromorphic Hardware",
    booktitle = "International Conference on Neuromorphic Computing Systems (ICONS)",
    doi = "10.1145/3589737.3605995",
    url = "https://doi.org/10.1145/3589737.3605995",
    publisher = "ACM",
    year = "2023"
}