the recognized MIT, Massachusetts Institute of Technology, announces the development of an artificial neuron that is, in his words, a million times faster than that of the human brain. “It is a faster computation for the artificial intelligence, with much less energy”, points out the organism.
Analog deep learning is a new area of artificial intelligence, which promises faster computation with a fraction of the power usage.
MIT engineers, using complex layered programmable repeating resistor arrays, created a network of artificial analog neurons and synapses. These execute calculations like a digital neural network.
The group of researchers used a practical inorganic material in the manufacturing process, which allows their devices “work a million times faster than previous versions, which is also about a million times faster than synapses in the human brain.”
MIT draws the parallels to better understand the discovery.
“In the human brain, learning occurs due to the strengthening and weakening of connections between neurons, called synapses. Deep neural networks have long adopted this strategy, where the weights of the network are programmed through training algorithms. In the case of this new processor, increasing and decreasing the electrical conductance of the proton resistors enables analog machine learning.”
— MIT
A Spaniard at the forefront of MIT research on the artificial neuron
Jesus A. del Alamo, Spanish engineer from MIT, from Department of Electrical Engineering and Computer Science, is at the forefront of the job.
“With this key information and the powerful nanofabrication techniques we have at MIT, we have been able to put these pieces together and show that the devices are inherently very fast, operating with reasonable voltages”, Del Álamo pointed out.
“This work has really put these devices to a point where they now look really promising for future applications.”
The working mechanism of the device is the electrochemical insertion of the smallest ion, the proton, in an insulating oxide to modulate its electronic conductivity.
Bilge Yidiz is part of the group developing the artificial neuron.
“Because we are working with very thin devices, we could speed up the movement of this ion by using a strong electric field, and bring these ionic devices to the nanosecond operating regime”, says Professor Yidiz.