Reservoir computing system developed to help machines to think just like humans
Scientists from University of Michigan in US have developed a new type of neural network chip using reservoir computing system to improve efficiency of teaching machines to think like humans.
The network developed using this system can predict words before they are said during conversation and help predict future outcomes based on the present.
The system has been inspired by brains, neural networks are composed of neurons, or nodes, and synapses, the connections between nodes. It was developed using memristors, a special type of resistive device that can both perform logic and store data.
The system used a special memristor that memorises events only in near history. This contrasts with typical computer systems, where processors perform logic separate from memory modules. These memristors required less space and can be integrated more easily into existing silicon-based electronics.
The system was proved to be functional using test of handwriting recognition, a common benchmark among neural networks. Only 88 memristors were used in the system as compared to conventional network that requires thousands for the task and achieved 91% accuracy.