Army laboratory invites professors to explore the future of computing

November 21, 2018

By RDECOM Research Laboratory Public Affairs

ADELPHI, Md. (Nov. 21, 2018) -- The RDECOM Research Laboratory, the U.S. Army's corporate laboratory (ARL), hosted a two-day event exploring the future of computing Nov. 13-14 at the Adelphi Laboratory Center.

Leading academics from top universities discussed neuromorphic computing. The concept mimics the operation of the human brain with artificial neurons and parallel operations.

"This looks beyond the current architecture that computers have used since the 1950s," said Dr. Madan Dubey, an Army researcher and one of the event organizers. "This is a very exciting area of research. The brain is powerful. In a traditional computer, you capture data, store it and process it. The brain performs all these tasks simultaneously and with extremely low energy."

In the Army lab's auditorium, the group listened to presentations about advances at different universities. Machine learning, deep learning, artificial intelligence, thinner chips, neural networks: speaker after speaker talked about advances that will propel computing into a new age.

Since the 1960s, Moore's law was an observation in the computer industry predicting the number of transistors in a dense integrated circuit would double about every two years. That pace has slowed, according to the event's first speaker, Harvard University Prof. David Brooks.

Brooks said under Moore's law, chip manufacturers should currently be unveiling products that would boost computing power by 80 times; however, a slowdown in technology cadence has resulted in an increase of only about 30 times.

With Moore's law coming to a possible finish line, researchers hope to achieve groundbreaking advances using new concepts. This may mark the beginning of the end for traditional computing, he said.

"There is a big question about the next technology," Brooks said. "I think this will open us up to new ideas and usher in a golden age of computing."

Brooks said discussing the future of computing at an Army lab was wholly appropriate. The Army has funded computer research since the 1940s, and that research has had an enduring impact, he said.

Mathematician John von Neumann (1903-1957) made major contributions to mathematics, physics, economics, computing and statistics. But, he is perhaps best known as a founding figure in computing for his work with the Army Ballistic Research Laboratory.

The von Neumann architecture outlines the need for a central processing unit, which includes a control unit and an arithmetic/logic unit. The CPU interacts with a memory unit and needs input and output devices.

Computing beyond the von Neumann model should include specialized accelerators, in-memory computing and non-digital components, Brooks said.

"Advances will require architects to recalibrate to go after big wins," Brooks said. "I think this is going to cause new ideas to percolate to the top."

In another presentation, Purdue University Prof. Kaushik Roy said the next generation of computing will offer tremendous potential for economic and societal impact.

"Current artificial intelligence apps are the tip of the iceberg," Roy said. "Neuromorphic computing will enable the next generation of autonomous systems."

The group discussed advances to bridge the gap from current computing capabilities to neuromorphic computing by using new algorithms and hardware.

One day a new computing paradigm may create machines that match and even surpass human intelligence and creativity. But researchers warned that there has been an overabundance of hype around the idea for decades.

For example, Brooks showed a 1958 New York Times article that said, "The Navy revealed the embryo of an electronic computer today that it expects will be able to walk, talk, see, write, reproduce itself and be conscious of its existence."

What is different today, Brooks said, is that there are commercially viable applications using deep learning and artificial intelligence.

"Neuromorphic computing is exciting," he said. "It's hard to imagine this is going to disappear."

Before the conference ended, officials from the lab's Army Research Office outlined research funding opportunities for individuals and university teams. The Army is still investing in future computing because all of the Army's modernization priorities rely on its success, officials said.

 

Last Update / Reviewed: November 21, 2018