Imagine yourself at a new AI-based startup company. Everyone else is busy with deep learning fads, but you approach your boss and suggest going in the direction of spiking neuron systems. Humoring you, your boss gives you your own standard everyday desktop computer without a GPU (thanks to the deep learning crowd, of course!) with a $300 allowance, and wants to see what kind of specialized hardware system you’re able to turn it into. What would you build, and what kind of performance increase would you expect to see?
In that case, I would look into “one-hot” representation for encoding spikes and I would make every bit to count. I.e. 1 byte allows you to encode neurons coming from 8 neurons. Small things that make the difference and I would try to use reliable frameworks such as caffe/tensorflow.
$300 covers proof of concept therefore don’t attempt to re-invent the wheel.
Bw
Pedro