Browsing by Author "MUNINDAR P. SINGH, Committee Member"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • No Thumbnail Available
    A Computational Model of Narrative Generation for Suspense
    (2007-05-18) Cheong, Yun Gyung; MUNINDAR P. SINGH, Committee Member; R. MICHAEL YOUNG, Committee Chair; BRADLEY S. MEHLENBACHER, Committee Member; JAMES C. LESTER, Committee Member
    The generation of stories by computers, with applications ranging from computer games to education and training, has been the focus of research by computational linguists and AI researchers since the early 1970s. Although several approaches have shown promise in their ability to generate narrative, there has been little research on the generation of stories that evoke specific cognitive and affective responses in their readers. The goal of this research is to develop a system that produces a narrative designed specifically to evoke a targeted degree of suspense, a significant contributor to the level of engagement experienced by users of interactive narrative systems. The system that I present takes as input a plan data structure representing the goals of a storyworld's characters and the actions they perform in pursuit of them. Adapting theories developed by cognitive psychologists, my system uses a plan-based model of narrative comprehension to determine the final content of the story in order to manipulate a reader's level of suspense in specific ways. In this thesis, I outline the various components of the system and describe an empirical evaluation that I used to determine the efficacy of my techniques. The evaluation provides strong support for the claim that the system is effective in generating suspenseful stories.
  • No Thumbnail Available
    Early Prediction of Student Goals and Affect in Narrative-Centered Learning Environments.
    (2008-08-14) Lee, Sunyoung; JOHN NIETFELD, Committee Member; MUNINDAR P. SINGH, Committee Member; R. MICHAEL YOUNG, Committee Member; JAMES C. LESTER, Committee Co-Chair; CARLA D. SAVAGE, Committee Co-Chair
    Recent years have seen a growing recognition of the role of goal and affect recognition in intelligent tutoring systems. Goal recognition is the task of inferring users' goals from a sequence of observations of their actions. Because of the uncertainty inherent in every facet of human computer interaction, goal recognition is challenging, particularly in contexts in which users can perform many actions in any order, as is the case with intelligent tutoring systems. Affect recognition is the task of identifying the emotional state of a user from a variety of physical cues, which are produced in response to affective changes in the individual. Accurately recognizing student goals and affect states could contribute to more effective and motivating interactions in intelligent tutoring systems. By exploiting knowledge of student goals and affect states, intelligent tutoring systems can dynamically modify their behavior to better support individual students. To create effective interactions in intelligent tutoring systems, goal and affect recognition models should satisfy two key requirements. First, because incorrectly predicted goals and affect states could significantly diminish the effectiveness of interactive systems, goal and affect recognition models should provide accurate predictions of user goals and affect states. When observations of users' activities become available, recognizers should make accurate "early" predictions. Second, goal and affect recognition models should be highly efficient so they can operate in real time. To address key issues, we present an inductive approach to recognizing student goals and affect states in intelligent tutoring systems by learning goals and affect recognition models. Our work focuses on goal and affect recognition in an important new class of intelligent tutoring systems, narrative-centered learning environments. We report the results of empirical studies of induced recognition models from observations of students' interactions in narrative-centered learning environments. Experimental results suggest that induced models can make accurate early predictions of student goals and affect states, and they are sufficiently efficient to meet the real-time performance requirements of interactive learning environments.