[4Yin2-02] Developing a Platform for Constructing Question Answering Systems Modeled after Real People
Keywords:dialogue system, deep learning
This paper proposes an application that can automatically build a dialogue system modeled on real people. This application film the user responding to questions and constructs a dialogue system that responds using the recorded video clips. The system uses the BERT model to select and play appropriate video clips in response to the user's input. Experimental results shows that the performance of the response selection model was improved than the performance of existing model. In the subjective experiment, our system received high scores in the categories of ``Personality'' and ``Consistency''
Authentication for paper PDF access
A password is required to view paper PDFs. If you are a registered participant, please log on the site from Participant Log In.
You could view the PDF with entering the PDF viewing password bellow.