Deep Learning

 

Kennesaw State graduates first Ph.D. in Data Science

KENNESAW, Ga. (May 15, 2019) — Designing new algorithms has been Linh Le’s focus for the past four years, and his hard work may someday impact applications for facial identification, stock market analysis and disease detection.

Le, the first graduate of Kennesaw State’s Ph.D. in Analytics and Data Science program, built an algorithm that helps to better classify data for use in decision-making, as part of his doctoral dissertation. He looked at students’ time spent studying and sleeping, visually graphing data and then further categorizing students by their GPAs.

His work in “Deep Embedding Kernel” integrates the strengths of deep learning and kernel methods, the two major branches in machine learning, and solves each of the branches’ individual weaknesses.

“I really enjoyed the research side of my program,” said Le, who completed both industrial and academic research as part of the Ph.D. course of study. “The program has helped me to develop my research expertise significantly, and I can successfully complete a wide range of data science tasks now.”

Linh Le

With a background in computing, Le enrolled in KSU’s Ph.D. in Analytics and Data Science – one of the first of its kind in the nation – in 2015. The four-year program is designed to train individuals to translate large, structured and unstructured, complex data into information to improve decision-making. Offered through KSU’s Analytics and Data Science Institute, the program is a multi-disciplinary degree that combines math, statistics and computer science.

Balancing doctoral-level courses and research opportunities, 29-year-old Le said that learning the fundamentals of data science and how analytics are used in research was demanding at the start of the program.  

“In the first two years, it was challenging to improve my math and statistics skills to a usable level for a Ph.D. program,” he said. “The first two years were quite stressful for me with all of the coursework, projects and publications I was doing.”

However, Le thrived as a scholar since applied research was a large component of the doctoral program.  

“Linh proposed a family of new machine learning methods that combines the strengths of two most significant branches of machine learning, namely deep learning and kernel machines, in one unified framework,” said Ying Xie, professor of information systems in the College of Computing and Software Engineering, who served as Le’s faculty mentor.

“This family of new machine-learning method has wide applications in the field of data science with superior performance, such as classification, dimension reduction, facial identification, stock market analysis and disease detection, to name a few,” Xie added.

Jennifer Priestley, executive director of the Analytics and Data Science Institute, associate dean of the Graduate College and professor of statistics, worked closely with Le to offer him critical feedback on his progress in understanding and utilizing data.

“Linh has been successful because he has treated his Ph.D. program like a job and comes in every day, whether he needs to or not,” she said.

“Our Ph.D. students have to understand not only the complex and emerging theory that is evolving in this nascent discipline, but they also have to understand how to apply what they have learned to multiple application domains ranging from fin-tech to health care to public policy. The ability to synthesize the theory and the application is an incredibly challenging intersection of skills,” she added.   

Le conducted large-scale experimental research alongside Xie in KSU’s Analytics and Data Science Institute’s Equifax Data Research Lab, a two-year-old sponsor in the Institute. Le investigated business challenges and opportunities created by the Atlanta-based Equifax’s non-traditional sources of consumer and commercial data.

Through their work with the Equifax Lab, Le and Xie currently have a patent pending for a “dual deep learning framework for big and unstructured data.” Le also contributed to multiple conference proceedings and presented at national and international forums for SAS, IEEE and the Association for Computing Machinery, on his research in deep learning, big data computing and data visualization.

Linh Le

According to Le, finding Kennesaw State’s doctoral program has been a major boon to his career path.

“I was actually Googling for a Ph.D. in Data Science at the end of my master’s program, and KSU came up as the only school which was offering such a program at that time,” said Le, who found an interest in basic analytical modeling and data warehousing while working on his master’s degree. Le has an engineering degree in Information Technology from the Hanoi University of Science Technology in Vietnam and his master’s degree in Information Systems from Marshall University in West Virginia.

During his time at Kennesaw State, he took advantage of every opportunity available to him, and was honored for multiple awards, including best paper in a research track at the Southern Data Science Conference and winner of the student poster competition for the SAS Analytics Experience.

With the distinction of being Kennesaw State’s first data science doctoral graduate, Le is looking forward to a future of more research studies, within industry or academia, and continuing to make new discoveries in the broad field of data science.

– Tiffany Capuano

Photos by Rob C. Witzel



A leader in innovative teaching and learning, Kennesaw State University offers more than 150 undergraduate, graduate and doctoral degrees to its more than 35,000 students. With 13 colleges on two metro Atlanta campuses, Kennesaw State is a member of the University System of Georgia and the third-largest university in the state. The university's vibrant campus culture, diverse population, strong global ties and entrepreneurial spirit draw students from throughout the region and from 92 countries across the globe. A Carnegie-designated doctoral institution, it is one of the 50 largest public institutions in the country. For more information, visit kennesaw.edu.

©