Artificial neural networks could make image processing and other artificial-intelligence techniques more powerful and practical and circuit can learn to recognize simple black-and-white patterns, thanks to devices called memristors located at each place the wires cross.

Memristors only existed in 2008 but prototype chip did not learn to do anything more difficult than recognize extremely simple black-and-white patterns. This was only useful if the computer was playing football against Newcastle, and anything designed with football in mind was not going to be the brightest of ideas.

But the idea of using chip circuits that mimic the neurons and synapses of biological brains is just getting started.

Dmitri Strukov, an assistant professor at the University of California, Santa Barbara told Technology Review that "neuromorphic" chips have been crippled by the fact that they use the same silicon transistors and digital circuits that make up ordinary computer processors.
Digital components are unsuited to mimicking synapses, many transistors and digital circuits are needed to represent a single synapse. By contrast, each of the 100 or so synapses on the UCSB chip is represented using only a single memristor.

Strukov has been writing in the pop science journal Nature about a new design which can be scaled for large networks
His group developed ways to control the process of making memristors so as to produce more reliable devices than had been made before, he says.

The UCSB group's simple chip is just a proof of concept, but the researchers believe their techniques can be scaled up to make larger, more powerful devices. Strukov says the technology could get a helping hand from the efforts companies such as HP and SK Hynix are making to commercialize memristors for data storage.