Title of Submission
Self-Annotation Through Conversational Grounding
Computer Science, MS
Major Advisor Name
Type of Submission
Current digital personal assistants are very good at performing routine tasks like setting up reminders etc. But, they do not attempt to establish common ground--the process of establishing mutual understanding--and require lots of initial data to improve their interaction process. In our work, we implement the incremental processing framework via a chatbot interface which not only grounds the system better, but the interactions provide labeled data to update the model and increase user-affinity towards such systems. Hence, subsequent interactions--post each update--bear semblance to a human-human interaction to a higher degree than such systems currently do.
Computer Science Department, Boise State University