Stefan Trausan-Matu's web page: http://www.racai.ro/~trausan
This talk will present some research for processing student online chat sessions recorded during collaborative mathematics problem solving in order to analyze and identify discourse structures. A corpus of Romanian language problem solving chats was developed. Statistical (Hidden Markov Models) and heuristic methods were used for identifying speech acts. In addition to computational linguistics methods, conversation analysis (Sacks, Garfinkel) is also used. This research was performed in connection with the NSF Virtual Math Teams (VMT) project at Drexel University, Philadelphia, PA, conducted by Professor Gerry Stahl. The VMT Project (http://mathforum.org/wiki/VMT) investigates Computer Supported Collaborative Learning (CSCL) in online collaborative math problem solving by small groups of students.
Prof. Stefan Trausan-Matu, PhD is a Fulbright Scholar Visiting Professor at Drexel University, Philadelphia, PA. He is professor at the Computer Science Department of the "Politehnica" University of Bucharest, Romania, where he teaches Natural Language Processing, Human-Computer Interaction, Cognitive Science, and Analysis of Algorithms. He is also a senior researcher at the Romanian Academy Institute for Artificial Intelligence (RACAI). He authored 10 books, 7 book chapters and more than 120 published papers and reports. He is vice president of the Romanian ACM-SIGCHI. His special research interests are: Computer-Supported Collaborative Learning (CSCL), Natural Language Processing (NLP), e-Learning, Human-Computer Interaction (HCI), Knowledge-Based Systems with special emphasis on ontologies, and philosophy.