The latest technology and digital news on the web

Human-centric AI news and analysis

How brain-like circuits could push accretion power to the next level

For the first time, my colleagues and I have built a single cyberbanking device that is able of artful the functions of neuron cells in a brain. We then affiliated 20 of them calm to accomplish a complicated calculation. This work shows that it is scientifically accessible to make an avant-garde computer that does not rely on transistors to account and that uses much less electrical power than today’s data centers.

Our research, which I began in 2004, was motivated by two questions. Can we build a single cyberbanking aspect – the agnate of a transistor or switch – that performs most of the known functions of neurons in a brain? If so, can we use it as a architecture block to build useful computers?

Neurons are very finely tuned, and so are cyberbanking elements that claiming them. I co-authored a analysis paper in 2013 that laid out in assumption what needed to be done. It took my aide Suhas Kumar and others five years of accurate analysis to get absolutely the right actual agreement and anatomy to aftermath the all-important acreage predicted from theory.

Kumar then went a major step added and built a ambit with 20 of these elements affiliated to one addition through a arrangement of accessories that can be programmed to have accurate capacitances, or abilities to store electric charge. He then mapped a algebraic botheration to the capacitances in the network, which accustomed him to use the device to find the band-aid to a small adaptation of a botheration that is important in a wide range of modern analytics.

The simple archetype we used was to look at the accessible mutations that have occurred in a family of bacilli by comparing pieces of their abiogenetic information.

Why it matters

The achievement of computers is rapidly extensive a limit because the size of the aboriginal transistor in chip circuits is now abutting 20 atoms wide. Any abate and the concrete attempt that actuate transistor behavior no longer apply. There is a high-stakes antagonism to see if addition can build a much better transistor, a method for stacking transistors or some other device that can accomplish the tasks that currently crave bags of transistors.

This quest is important because people have become used to the exponential advance of accretion accommodation and ability of the past 40 years, and many business models and our abridgement have been built on this expectation. Engineers and computer scientists have now complete machines that aggregate astronomic amounts of data, which is the ore from which the most admired commodity, information, is refined. The volume of that data is almost acceleration every year, which is outstripping the adequacy of today’s computers to assay it.

What other analysis is being done in this field

The axiological theory of neuron action was first proposed by Alan Hodgkin and Andrew Huxley about 70 years ago, and it is still in use today. It is very circuitous and difficult to simulate on a computer, and only afresh has it been reanalyzed and cast in the mathematics of modern nonlinear dynamics theory by Leon Chua.

I was aggressive by this work and have spent much of the past 10 years acquirements the all-important math and addition out how to build a real cyberbanking device that works as the theory predicts.

There are abundant analysis teams around the world taking altered approaches to architecture brainlike, or neuromorphic, computer chips.

What’s next

The abstruse claiming now is to scale up our proof-of-principles affirmation to article that can attempt adjoin today’s agenda behemoths.

Published October 11, 2020 — 18:00 UTC

Hottest related news