ChiQat-Tutor: A Natural Language Processing Component That Uses Artificial Intelligence to Interpret and Answer Student Queries
Session Number
F15
Advisor(s)
Barbara Di Eugenio, University of Illinois at Chicago
Location
A-121
Start Date
28-4-2016 9:15 AM
End Date
28-4-2016 9:40 AM
Abstract
Intelligent Tutoring Systems are computer systems that are designed to provide users with customized or immediate instruction without the use of a human teacher. An example of this is ChiQat-Tutor, which is a system designed to teach students basic Computer Science concepts. The purpose of this research was to develop and evaluate a component of ChiQat-Tutor that interprets and answers students’ questions regarding the concepts that the system teaches. My component uses Artificial Intelligence software that analyzes a user’s query and code to generate an appropriate and accurate response. In order to create this component, I first analyzed data from student-tutor dialogues that had been collected earlier to understand the types of questions students ask when learning Computer Science. I then created a keyword-based system, programmed in Java that interprets and answers students’ queries by analyzing their code and query in order to select and return the best possible response to their question. After creating this component, I conducted an evaluation of it. Per the results, 79.2% of the time, the question-answering component of ChiQat answered the user’s question in an appropriate and accurate manner. Thus, I can conclude that using a keyword-based, Artificially Intelligent question-answering system can provide both apt and precise responses to student inquiries, moving Intelligent Tutoring Systems one step closer to simulating a live student-to-tutor experience.
ChiQat-Tutor: A Natural Language Processing Component That Uses Artificial Intelligence to Interpret and Answer Student Queries
A-121
Intelligent Tutoring Systems are computer systems that are designed to provide users with customized or immediate instruction without the use of a human teacher. An example of this is ChiQat-Tutor, which is a system designed to teach students basic Computer Science concepts. The purpose of this research was to develop and evaluate a component of ChiQat-Tutor that interprets and answers students’ questions regarding the concepts that the system teaches. My component uses Artificial Intelligence software that analyzes a user’s query and code to generate an appropriate and accurate response. In order to create this component, I first analyzed data from student-tutor dialogues that had been collected earlier to understand the types of questions students ask when learning Computer Science. I then created a keyword-based system, programmed in Java that interprets and answers students’ queries by analyzing their code and query in order to select and return the best possible response to their question. After creating this component, I conducted an evaluation of it. Per the results, 79.2% of the time, the question-answering component of ChiQat answered the user’s question in an appropriate and accurate manner. Thus, I can conclude that using a keyword-based, Artificially Intelligent question-answering system can provide both apt and precise responses to student inquiries, moving Intelligent Tutoring Systems one step closer to simulating a live student-to-tutor experience.