System Bits: April 21

What DARPA is investigating; MIT’s tech forecasting model.


DARPA’s Research
DARPA’s Semiconductor Technology Advanced Research Network, aka Starnet, unveiled its research plans for 2015 and 2016. Topping the list in 2015 is an investigation into the feasibility of using advanced 2D materials for ultra low-power devices, along with the fab methodology, modeling and simulation tools necessary to make it all work.

The fiscal 2015 research will look at limits in scaling multifunctional and spintronics materials, as well as begin development of new heterogeneous architectures and the statistical underpinnings of machine learning and neuro-based architectures, as well as developing components, architectures, data control and sensors for swarm applications.

Next year, the agency plans to design circuits using “steep turn-on” transistors for pattern recognition and energy scavenging, along with developing scalable computing system concepts for the 2020 to 2030 timeframe for defense-related computing systems.


Forecasting Change
MIT engineers have developed a formula for predicting how fast certain technologies are moving based upon patent information.

Researchers scanned more than 500,000 patents spanning everything from solar photovoltaics to fuel cells to 3D printing. One surprise was that just because there are more patents for a particular technology doesn’t mean that it’s actually progressing faster. That’s evident with 3D printing, which has fewer patents than other technologies, but it’s progressing at the speed of Moore’s Law.

First of all, the fastest-moving technologies are 3D printing, MRI technology and wireless communications.

“There’s a lot of nuance to our method, and I don’t see it as something to hand out to the masses to play with,” said Chris Benson, a former Department of Mechanical Engineering graduate student at MIT, on the school’s Web site. “I see it more as something where we work with somebody to help them understand what the future technological capabilities that they’re interested in are. We’re probably more like a real estate agent, and less like Zillow.”

Researchers believe the method may evolve into a rating system, similar to some of the stock market indices such as Standard & Poor’s, which may help investors find the next big thing and avoid investments that are unlikely to succeed.