A University of Texas at Dallas mathematician has received a five-year grant from the National Science Foundation (NSF) in support of her work on doing more with less data.

Dr. Yifei Lou

Dr. Yifei Lou, assistant professor of mathematical sciences in the School of Natural Sciences and Mathematics, was awarded an NSF Faculty Early Career Development (CAREER) Award of more than $400,000 for her work on “Mathematical Modeling for Data to Insights and Beyond,” a project that seeks analytical tools to provide guidance on acquiring data more efficiently. In an age when so much focus is on big data, she calls her work “small data.”

“If one is able to collect only a certain amount of data, what method will allow him or her to get the most out of it?” Lou asked. “For example, if you can see only 25 percent of the pixels in an image, what way of picking one out of every four pixels would let us best identify the pictured object?”

Although big data is widely used by those who have the means, Lou said it remains out of reach for many fields and techniques for various reasons, many of which are technical or economic.

“Take computed tomography (CT) scanning, for example — the scans we undergo for medical reasons,” she said. “The maximum safe radiation dosage limits the amount of data that can be recorded, so how we reconstruct the image of your body has to be aided by prior knowledge.”

That prior knowledge in Lou’s theory involves incorporating what she calls “additional information from the application domain” — simply put, context that tells you what features your scan will not need to determine.

“What is the best way to scan when you have a maximum you can’t exceed?” Lou asked. “In the case of a CT scan, a body has elements that are what we call ‘piecewise constant’ — there will be one heart, one stomach, and their position relative to each other is very regular. So rather than working with an infinite number of possibilities, we can teach the computer which elements it doesn’t need to worry about.”

About CAREER Awards

The Faculty Early Career Development Program supports early-career faculty who exemplify the role of teacher-scholars through outstanding research and excellent education. The highly selective program is the National Science Foundation’s most prestigious award for junior faculty who are considered likely to become leaders in their fields.

Another example Lou cites is seismic data — a field in which data gathering is prohibitive due to the scale of information that can be required.

“If I can devise a strategy that dictates what points of view should be recorded, then it can save time, money and computational resources,” she said.

Similar logic could be applied to other data-intensive fields, such as facial recognition, genomics and astronomy, although Lou is not specifically focusing on these areas.

“As an applied mathematician, I’m looking for ways to play a role in developing technology that makes a difference in everyday applications,” she said. “Hopefully, our guidance and computational tools will lead scientists to use better methods of data collection.”

Dr. Vladimir Dragovic, the department head of mathematical sciences, described Lou’s grant as recognition of the high quality of research produced by his department and across the School of Natural Sciences and Mathematics.

“Dr. Lou has received this most prestigious NSF award for junior faculty for a project that is deeply integrated in our departmental efforts, in particular for the development of data science programs,” he said. “Together with colleague Dr. Tomoki Ohsawa [assistant professor of mathematical sciences], she has attracted many undergraduates to mathematical modeling. This award is an inspiration for others in our quest to become a top-ranked math department at a national research university.”