Research Scientist in the Center for Educational Informatics at North Carolina State University
Jonathan Rowe is a Research Scientist in the Center for Educational Informatics at North Carolina State University, as well as an Adjunct Assistant Professor in Computer Science. He received his Ph.D. in 2013 from North Carolina State University, and his B.S. in 2006 from Lafayette College. His research investigates artificial intelligence and human-computer interaction in game-based learning environments, with an emphasis on interactive narrative technologies, educational data mining, intelligent tutoring systems, user modeling, and player engagement. His work has been recognized with best paper awards, including best paper at the Seventh AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment and best paper at the Second International Conference on Intelligent Technologies for Interactive Entertainment. He has served as the technical lead on several game-based learning projects, including Crystal Island: Lost Investigation, which was a finalist for Best Serious Game at the 2012 Unity Game Awards, as well as the 2012 I/ITSEC Serious Games Showcase and Challenge. He serves on the editorial board of the International Journal of Artificial Intelligence in Education (IJAIED), and he is Program Chair for the 2017 AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE-17).
Learning with Games that Learn: Personalizing Serious Games with Machine Learning
Recent years have seen immense interest in machine learning for a vast array of applications, including games. Machine learning techniques enable models that recognize patterns in player behavior, models that draw inferences about player engagement, and run-time AI systems for dynamically personalizing gameplay. This talk will provide an overview of work in the Center for Educational Informatics at North Carolina State University investigating applications of machine learning to player gameplay data in several serious games for science education. The talk will focus on two primary modeling tasks. First, it will describe applications of deep learning, such as long short-term memory networks, for player goal recognition in open-world games. Second, it will describe how reinforcement learning techniques can induce models for tailoring game events to individual players. The talk will conclude with a discussion of future directions on how machine learning can help us to improve our understanding of players, as well as to craft better gameplay experiences.