While "big data" has become an industry buzzword lately, NASA Langley Research Center has been dealing with it for decades, said Manjula Ambur, head of NASA Langley's Information Management Branch. Technologies are just maturing to analyze data in a deeper way.
"It's more about big analytics," Ambur said. "Really the power of analytics to provide us insight and predictions from our data and information for decision-making and innovation."
New data frontier
Over the years, researchers, government, academia and commerce have been collecting and storing data. That information has been used in modeling and simulation, experiments and research and development. Now, a new data frontier awaits — one where computer scientists can use it to develop machines that think and problem-solve like humans.At NASA Langley, the Big Data Analytics team envisions computer assistants or "virtual helpers" doing the more mundane work for scientists, who can then concentrate on bigger projects or ideas. Ultimately, Ambur, who is associate chief information officer of Big Data Analytics, anticipates this could lead to even more scientific discoveries in less time.
Since the Big Data Analytics and Machine Intelligence effort launched at NASA Langley about a year and a half ago, the team has been working with researchers and engineers to figure out how to apply it to different disciplines, like aeronautics.
'Virtual helpers'
One goal is to create "virtual helpers," which could be robots, high-tech watches or smartphone apps. Who knows how they'll be manifested, Ambur said. NASA Langley has been testing about a dozen applications in pilot projects over the last seven months.This is using the same technology behind "Watson," a computer system developed by IBM researchers to process natural language and stored information to answer questions posed on Jeopardy! in 2011. Watson, which won $1 million on the show, could "learn" from human interaction and by adapting to new information. IBM has since improved the Watson technology and opened it up to businesses for "cognitive app" development and partnered with USAA last year to develop an app to help military personnel transition to civilian life.
So how can these digital assistants help NASA Langley? Scientists currently look at various CT scans of materials like stainless steel or blended composites to find anomalies so that they can tweak their compositions to make them safer. A machine using trusted algorithms could scan more images faster to narrow down what images scientists should look at, or it could detect patterns that scientists may not see, Ambur explained.
This would make it easier to test the quality of 3-D printed objects, particularly as 3-D printing — forming an object layer by layer using plastic or other feedstock — has been identified as a way to manufacture parts in space, she added.
Virtual helpers can also analyze incoming data to alert scientists when an expensive model aircraft may be about to break in wind tunnel tests, Ambur said. Obviously, preventing their destruction would save time and money.
Machines capable of deep analytics wouldn't replace scientists but would empower them to focus on higher-level research, she said. Machines don't have biases or get tired like people, and could offer their own solutions, too.
"Imagine if Siri can answer these kinds of questions," Ambur said, referencing iPhone's personal assistant app.
After NASA Langley learns how to apply big data analytics in various case studies — and validate that it works — the Big Data Analytics team would like to come up with a suite of internal tools that could be used or customized for various projects, Ambur said.
Potential in other fields
While NASA's big data analytics is tied mostly to physics and a scientific discipline, Ambur sees great potential for its use in many fields, including social media, finance and retail. In fact, the creation of such intelligent machines or software could empower nonscientists or nonexperts to come up with innovations, she said.Dr. Charles Frazier, chief medical information and innovation officer at Riverside Health System, agreed. Imagine Watson-type helpers taking a patient's health records, doctor's notes and lab results and then processing that along with decades of health research to propose possible diagnoses to a doctor, he said. That type of data analysis may also be useful in prescribing drugs or tailoring treatments to a patient's genetic profile.
"I think it's an amazing time in medicine," Frazier said.
Veteran-led EMB Information Technologies at the Peninsula Technology Incubator in Hampton would like to help companies make use of their own data by customizing already-built software to keep it affordable.
The startup, which is an IBM partner, sees growth opportunities in cybersecurity, finance and health care in particular. Right now, the company is partnering with a couple other small businesses in Hampton to demonstrate their capabilities. EMB President Edward Ballanco and Vice President of Technology Chris Schlitz both come from the modeling and simulation world.
"So many companies have significant amounts of data and they're not using it," Schlitz said. "We think it's the next big thing."
Between the military, defense contractors, Jefferson Lab, NASA Langley and local universities, Hampton Roads has a good bit of big data activity going on, Spectrum Chief Scientist Frank Byrum said. That means the workforce is there to beef up data science work to the commercial sector, he added.
Byrum, who works for Newport News-based defense contractor Spectrum in cybersecurity, knows that as defense dollars shrink, local information technology companies will be looking for other customers even if it's not technically big data. Still, he understands privacy, data storage and data ownership issues will need to be addressed as data science takes hold of various industries.
"Many organizations can benefit from taking some of the ideas from big data and looking at the way they actually use information in their business," Byrum said. "... If you're a teenager looking for a good field to go into, go into data science."
©2015 the Daily Press (Newport News, Va.)
