Professor’s sabbatical yields virtual adviser prototype

RANKINEN

RANKINEN

During his sabbatical last spring, Jeff L. Rankinen planned to investigate the possibilities of artificial intelligence. The Pennsylvania College of Technology associate professor returned to the classroom this fall after transforming one of those possibilities into a reality.

Rankinen was part of a four-person team that earned $15,000 in funding to develop a “virtual adviser” as part of the Penn State EdTech Network’s Nittany Watson Challenge, which tasked entrants to improve the student experience via artificial intelligence. Just five of 39 teams were awarded $15,000 to create both a working prototype and minimum viable product in conjunction with IBM Watson, a technology platform focused on data analysis, natural language processing and machine learning.

“I thought the competition would be a good opportunity to learn more,” said Rankinen, who has taught electronics and computer engineering technology at Penn College since 1986. “I have been interested in artificial intelligence since beginning my graduate work at Penn State in 1987. It was very enriching to get immersed in artificial intelligence with IBM Watson during the competition.”

That immersion yielded “AI Assisted PSU IST Helper,” also known as “Leo,” a virtual adviser meant to assist both students and faculty. The system is programmed to answer common student questions directed at faculty advisers and facilitate appointment scheduling if it can’t adequately respond to a query. Available any time of day, Leo grows smarter with each question asked as it stores additional data.

Rankinen’s team consisted of two Penn State students who have since graduated: Michael Morelli, information sciences and technology, and Parantap Tripathi, mechanical engineering, as well as Matt Dalesio, business intelligence developer for Penn State World Campus. Dalesio earned a bachelor’s degree in electronics engineering technology (now electronics and computer engineering technology) in 2008 from Penn College.

The group worked with three IBM mentors in developing the minimal viable product, or beta form of the system.

“We were able to do some student testing and were encouraged that 85 percent of the students tested felt that it was useful,” said Rankinen, who devoted 300-plus hours to the project. “The sample size was 80. Our hope is to further develop it and do a pilot test. The goal is to see it be developed for use at Penn State and Penn College.”

“Jeff’s impressive work for the competition is inspiring,” said David R. Cotner, dean of industrial, computing and engineering technologies. “His tenacity in stretching the boundaries of artificial intelligence shows how dedicated he is to his field and to his students. I know Jeff will find ways to incorporate his experiences to benefit the students in his classroom and lab.”

One student has already benefitted. Yahya S. Rumaili, of Williamsport, a plastics and polymer engineering technology major, designed and produced, via a 3-D printer, the plastic enclosure for the virtual adviser.

Rankinen’s students enjoy access to IBM Bluemix, a cloud-based computing system offering artificial intelligence services. He also is devising a project for students to create a text-based chat bot capable of answering questions.

“My plan is to involve each of my classes with artificial intelligence to some degree,” he said.

Rankinen holds a doctorate and master’s degree in materials engineering from Penn State, a bachelor’s degree in electrical engineering from Rochester Institute of Technology, and associate degrees in electronics technology and computer information systems from Penn College’s immediate predecessor, Williamsport Area Community College. He is co-author of the textbook “Introduction to Engineering Technology.”

For information on the Nittany Watson Challenge, visit http://edtechnetwork.psu.edu/watson.

Penn College is a national leader in applied technology education and workforce development. Visit www.pct.edu, email admissions@pct.edu or call toll-free 800-367-9222.

COMMENTS