Tiny Neural Network - 3 Spiking Neurons

Small artificial Neural Network consisting of 2 input neurons and 1 output neuron. All three neurons are spiking neurons. Network is configured to resemble logical AND element. Input neurons are on the left side of the circuit and are tracked with orange and red colors. The output neuron is tracked green. Two potentiometers in the center represent connection weights (SYNAPSES aka neuron connections) from input to the output neuron. If switch next to one of the input neurons is closed - constant current is injected and input neurons start firing and sending excitatory signals (SPIKES) via potentiometers (SYNAPSES) to the output neuron. Initially network is configured to resemble logical AND element - output neuron will fire only when both input neurons are activate and their spikes roughly coincide in time. Higher resistance of potentiometer will narrow spike correlation window. In contrast by changing potentiometers values to 25% will turn network into logical OR. Network logical layout: (orange)-. \ (green)-->(V) / (red)--' SEARCH for "Spiking Neuron" circuit for more details about internals of a single neuron.

published 9 years ago