Electronic Theses and Dissertations
INTEGRATION AND EVALUATION OF A NEW NLP METHOD TO ENHANCE THE PERFORMANCE OF CONVERSATION BASED ADAPTIVE INSTRUCTIONAL SYSTEM (CBAIS)
Date of Award
Master of Science
Electrical and Computer Engr
The proposed Master thesis is to develop, implement, and evaluate a new method that dramatically increases the efficiency and accuracy of conversation-based adaptive instruction systems (CbAISs). The outcome of the work will directly serve the ElectronixTutor (sponsored by the US Navy) project. The method developed will be applicable to Conversation-based AISs universally and domain-free.The work is guided by the cognitive principle of human learning. It is proven that one-to-one teaching achieves much better learning gain compared with teaching in traditional classrooms. The observed advantages of one-on-one teaching are largely due to the fact that it is adaptive and individualized. After investigating hundreds of hours of teacher-student interactions, researchers found an interesting fact that teachers (either experts or novice teachers) could achieve significant learning by applying simple pedagogical strategies called EMT (expectation-misconception tailored) dialog. EMT is the foundation for a prototypical CbAIS (Adaptive Instruction System) developed at the University of Memphis. The proposed method in this thesis work is to maximize the performance of EMT applications used in CbAISs.The computer engineering aspect of the thesis project is to demonstrate a method that integrates a suite of technologies in a service-oriented architecture (SOA). Specifically, the thesis project makes use of cloud computing (such as Google VM, or AWS), server-side scripting such as NodeJS. The services are in the form of RESTful API to achieve flexible and fast deployment while keeping the enhanced CbAIS scalable.The application is successfully integrated into the Authoring and quality control of the ElectronixTutor. It is expected that the time consumed in the authoring and quality control process will be dramatically reduced. The methods developed in the thesis will have a positive impact on the wide adoption of CbAIS as an advanced learning environment that can be used to help students learn in other domains, such as STEM majors.
Dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.
Wang, Lijia, "INTEGRATION AND EVALUATION OF A NEW NLP METHOD TO ENHANCE THE PERFORMANCE OF CONVERSATION BASED ADAPTIVE INSTRUCTIONAL SYSTEM (CBAIS)" (2021). Electronic Theses and Dissertations. 2362.
Data is provided by the student.