Haiying Li: Real-Time Assessment and Scaffolding of Adaptive Learning @ Pless Hall, 7th Floor, ALT Conference Room 785

02/15/2018 11:00 am – 02/15/2018 12:00 pm
Real-Time Assessment and Scaffolding of Adaptive Learning
Haiying Li, Rutgers University
Thursday, February 15, 11:00AM-12:00PM
Pless Hall, 7th Floor, ALT Conference Room 785
Recently, the majority of efforts toward developing real-time assessment in online learning environments have focused on students’ behaviors and actions, despite the fact that students’ written responses may reveal rich information about the challenges and difficulties they confront during learning. One big challenge with written data is that they are much more coarse relative to click stream data. In this talk, Haiying will present three automated assessments for: (1) collaborative problem solving in an educational game, (2) scientific explanations in a virtual, intelligent tutoring system (ITS), and (3) summaries in a conversation-based ITS. In the first study, automated classifiers of speech acts, question categories, and personalities were developed to identify players’ challenges and social skills in order to provide individualized support during collaborative learning. In the second study, pattern-algorithms were used to automatically score scientific explanations in the format of claim, evidence, and reasoning. This fully operationalized assessment enables the development and implementation of real-time, adaptive scaffolding. The third study presents a crowdsourcing-based latent semantic analysis to automatically and economically score summaries. The findings across these three studies indicate that real-time assessment using techniques of computational linguistics and machine learning supports scaffolding of student performance in adaptive, online learning, and assessment environments.
Haiying Li is a postdoc in the Graduate School of Education at Rutgers University. She holds a Ph.D. in Cognitive Psychology from the Department of Psychology and the Institute for Intelligent Systems at the University of Memphis, and a M.A. in English Linguistics and a B.A. in English Education both from Hebei Normal University in China. Haiying’s research is at the intersection of cognitive science, learning sciences, computational linguistics, and data science. She has designed adaptive intelligent systems that support reading comprehension and science inquiry practices. To make online systems more adaptive, she has developed automated assessments for summaries, scientific explanations, and collaborative problem solving. Haiying contributes to learning analytics research by expanding from log file analytics to language and discourse processing, which captures invaluable information for improving teaching, learning, and assessment. She has also developed automated text analysis tools to support discourse processing, including Coh-Metrix, a measure of formality, and topic modeling. She has about 100 publications and presentations in journals, book chapters, and refereed conference proceedings. Her research on network-based teaching won the Excellent Academic Paper Award and the Award for Excellent Research from the Beijing Higher Education Association, China. Her paper on measures of formality won the John Castellan Award for Best Student Paper at the Society for Computers in Psychology.