Building functions into microchips mean they run fast and efficiently, despite the fact it reduces the flexibility of what the hardware can do.
Artificial neural networks can require a lot of computer processing power because every simulated neuron in the network is essentially a mathematical procedure that needs running every time the network is updated.
With a frequently updated network of thousands and thousands of simulated neurons, the required computing power quickly adds up.
What Boahen and others are doing is building microchips that have functions to simulate neurons built into the hardware to make this possible on only a few chips.
Crucially, some of the simulation is done by analogue, rather than digital, computation.
Digital processors use transistors in their on/off switching phase. Mead realised that by using transistors in their analogue amplifier phase instead, he could build circuits that accurately mimicked the electrical behaviour of real neurons. Using transistors in this way also meant Mead could dispense with the central clock altogether, dramatically cutting the power demand. As long as the input signals arrived within a few milliseconds of each other, the circuit of transistors imitating a given neuron would sum the input values, and if that tipped over a certain threshold, would produce an output spike. “It’s totally foreign to the way we’ve built computers for the last 40 years,” says Boahen.
Neural networks can try and simulate real neurons as closely as possible, or be quite abstract or general impressions of them.
The New Scientist article notes that these hardware-based systems are also intended to mirror the brain’s biology quite closely.
Closely enough, that the systems are being used for designing replacement retinas to augment parts of the damaged visual system in humans.
Unfortunately, the article isn’t freely available online, but you should be able to pick up a copy in your local library or newsagent.
However, the Brains in Silicon lab at Stanford University has a comprehensive website with a host of information if you want to find out more.