A Virtual Dialogue Assistant for Conducting Remote Exams

Abstract

In this paper, we demonstrate issues and possible solutions to building an Artificial Intelligence Dialogue Assistant for human-machine communication. We specialize it for conducting written exams at online education platforms, talk about the main logical components of the system: knowledge base, question encoder, question generation module, question analysis module. As a knowledge base we consider text fragments representing parts of the course representing parts of a course and is a source for a format ontology; and is also sourced for neural network generation of fact-based questions in question generation module and building dependency trees for answer evaluation in question analysis module.

Keywords

• speech technologies
• artificial intelligence
• human-machine communication
• natural language processing
• linguistic
• deep learning