Monday, November 28, 2016

Neuromorphic Computing


Engineers have been developing microprocessors which mimic the biological synapses of the human brain. The projection is being led by Joshua Yang and Qiangfei Xia, professors of electrical and computer engineering at the University of Massachusetts. Their work focuses on ionics instead of electronics, specifically “memristors,” which “enable neuromorphic computing by reproducing the functions in biological synapses and neurons in a neural network system, while providing advantages in energy and size.” These synaptic emulators have a distinct advantage over traditional microprocessors because they are not dependent on a power source. Essentially, a memristor has a memory, so even if it loses power it can remember what it was doing before and continue the action. Computers of the future may be able to shut on and off as fast as a light bulb without losing any data, files, or processes. What implications might this technology have upon our future?

No comments:

Post a Comment