Personalized ECA Tutoring with Self-Adjusted POMDP Policies and User Clustering
Abstract
An Embodied Conversational Agent (ECA) is an intelligent agent that enables real-time human/computer interaction in natural language. For its rich style of communication, ECA is particularly popular and useful in applications such as education, e-commerce, healthcare, finance, marketing, and business, where a human-like conversation is more attractive to users than traditional keyboard-based interaction. The interest in using ECA in e-learning has become even stronger since the COVID-19 outbreak, and a preliminary investigation has been started by our research group to extend collaborative learning in a virtual environment with personalized ECA tutoring.
This thesis document first highlights the prior work of personalized tutoring with ECA, including wavelet transformation for user clustering and face-to-face interaction for quiz-style e-learning. An enhanced approach is then developed to enable self-adjustment of POMDP policies for dialogue management and to allow a more natural way of question/answer style of personalized tutoring with a generic, flexible tutoring ontology. In addition, the proposed approach uses machine learning techniques to adjust knowledge levels of user clustering and evaluates its effectiveness by conducting experiments with real datasets. This research work is projected to further improve online learning with ECA serving as a personal tutor.