Knowledge graph based question answering system on COVID-19 for students

• Main Author: Chen, Ping
• Other Authors: Miao Chun Yan
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
• Online Access: https://hdl.handle.net/10356/144609
• Institution: Nanyang Technological University

With technological advancement, people enjoy the convenience of searching for answers to their questions directly from the Internet, instead of traditional ways like searching for answers from books. COVID-19, a highly contagious disease, has caused the death of more than one million people globally since December 2019. During the pandemic, people tend to have the desire to learn more about the disease. However, due to the large amount of information online, people may not always get what they want immediately. It may take time for them to find the information that they really want to know. Meanwhile, to most people, more is unknown than known. They may receive fake information without noticing it. A bilingual knowledge graph based question answering system was developed for students of different educational background. A knowledge graph was built from an open-source dataset to store relevant information. Via semantic parsing, the system is able to identify the class that a user input question belongs to. Based on the class label, an answer can be retrieved from the knowledge graph via a query language. The proposed solution provides students a quick, natural, and intuitive way of acquiring knowledge in a language that they are comfortable with.