Beyond Artificial Intelligence, WVU researchers pursue advances in deep learning

MORGANTOWN — West Virginia University's Benjamin M. Statler College of Engineering and Mineral Resources pursues research in the field of deep learning that can have impact in the fields of medicine and forensics sciences.

Thirimachos Bourlai is an associate professor with the college's Lane Department of Computer Science and Electrical Engineering. He said the concept of deep learning falls under the umbrella of machine learning and artificial intelligence (AI). But deep learning, he said, involves automated decision making and data processing on a larger scale using more complicated algorithms and layered networks.

"Basically, in the past, AI has been considered science fiction but it's also the future," Bourlai said. "We've seen movies like 'The Terminator' and it's still a very cool movie to watch, but it (AI) has become part of our everyday lives."

Bourlai said the concept of deep learning was only coined in 2012 but is expected to have a revolutionary impact on society and industry as great as Benjamin Franklin's early experiments with electricity 270 years ago.

The ultimate goal of WVU's research, along with those of other universities and companies, is to foster better decision making by programs. The WVU Multispectral Imagery Lab, which Bourlai founded in 2010, is already researching applications for deep learning in the field of biomedical technology.

One example that came out in 2016, Bourlai explained, was an algorithm that automatically sorts images of patients left and right irises, going so far as to determine the left and right iris for an individual patient out of a thousand, something that can come in handy if images are mislabeled.

"This is a critical concept," he said. "You can do the same with other mislabeled data. With deep learning, if you have a lot of information, the system can learn and understand and get better."

Bourlai said other deep learning systems under development at WVU is the detection of irregular heart beats. As for deep learning applications throughout the world, he said this technology can recognize and identify a single individual among one million in a database with just their photo faster than any human could.. This is something, he said, that could help hospitals and law enforcement, adding that while the exact applications have yet to be determined, deep learning could help tackle West Virginia's opioid crisis.

Deep learning is also at the heart of autonomous vehicles, allowing them to recognize pedestrians, traffic signals along with other vehicles and how fast they're going. Bourlai said there are still limitations to overcome. These vehicles can recognize potholes but humans still maintain the best judgement when determining if it's safe to actually go around it.

Anne Barth serves at the executive director of TechConnectWV, an economic development organization that focuses on fostering advanced energy, biometrics, biosciences and chemicals.

She said AI and deep learning will most likely come to play a role in the West Virginia's energy sector. Continued research into this field will also encourage tech companies to set up shop in West Virginia. What's more, thanks to the gig economy Barth said work forces aren't as concentrated as they were in the past and anyone with an internet connection can participate in this growing market.

"What's been a revelation for me to understand is that you can go and get training in software coding and earn a very good income after maybe a 14 or multi-week boot camp," Barth said. "In software programming there is another avenue for workforce opportunity, and I'm very excited about this for West Virginia."

She said as these camps spring up throughout the state, more companies across the nation are engaging in 'here-shoring' to bring back back software jobs off-shored to other countries.

"We can be competitive, and its a fabulous opportunity for a state like West Virginia," she said, noting that technology jobs can be a new source of income for those displaced by the decline of the coal industry.

Staff writer Conor Griffith can be reached by at 304-395-3168 or by email at cgriffith@statejournal.com



SHARE THIS
Previous Post
Next Post