Summary
International Symposium on Nonlinear Theory and Its Applications
2016
Session Number:A3L-G
Session:
Number:A3L-G-2
Two Heuristic Approaches to Parameter Tuning for an Analog Silicon Neuron Circuit
Ethan Joseph Green, Takashi Kohno,
pp.-
Publication Date:2016/11/27
Online ISSN:2188-5079
DOI:10.34385/proc.48.A3L-G-2
PDF download (476.4KB)
Summary:
Neurons play a major role in memory, cognition, sensory processing, body regulation, and a host of other functions vital to organisms. Analog silicon neurons are biologically inspired VLSI (very-large-scale integrated) circuits that mimic the electrophysiological behavior of neurons. This research looks at circuit parameter tuning for an ultra-low power analog silicon neuron designed with qualitative neuronal modeling. A key challenge to operating this circuit is adjustment of the circuit parameters to allow for similar behavior across a range of temperatures and eventually amongst many silicon neuron circuits in a silicon neuronal network. Two heuristic approaches were applied to the silicon neuron to supplement trial-and-error-based tuning of the circuit's parameter voltages. In the future, these two approaches will be combined to create a fully automated tuning algorithm.