Researchers at IBM have designed a chip, called TrueNorth, that is modeled on the neuronal networks that make up the brain. The chip is comprised of 5.4 billion transistors which are organized into one million "neurons", contrary to the organization of conventional chips, in which the transistors are arranged in binary "Yes/No" logic gates.
Instead of a strictly binary decision of "yes" or "no", these neurons are able to take input from several other neurons, firing a signal to the other neurons it is connected to once its inputs reach a certain threshold level. With the concurrent interconnected firing of millions of these neurons, the chips can begin to recognize patterns of inputs, such as those in "vision, audition, and multi-sensory fusion, and [they] has the potential to revolutionize the computer industry by integrating brain-like capability into devices where computation is constrained by power and speed. These systems can efficiently process high-dimensional, noisy sensory data in real time, while consuming orders of magnitude less power than conventional computer architectures."