Dynamic causality knowledge graph generation for supporting the chatbot healthcare system
Authors
Yu, Hong QingAffiliation
University of BedfordshireIssue Date
2020-10-31Subjects
chatbotknowledge graph
causality analysis
natural language processing
artificial intelligent
healthcare
Metadata
Show full item recordOther Titles
Proceedings of the Future Technologies Conference (FTC) 2020, Volume 3Abstract
With recent viruses across the world affecting millions and millions of people, the self-healthcare information systems show an important role in helping individuals to understand the risks, self-assessment, and self-educating to avoid being affected. In addition, self-healthcare information systems can perform more interactive tasks to effectively assist the treatment process and health condition management. Currently, the technologies used in such kind of systems are mostly based on text crawling from website resources such as text-searching and blog-based crowdsourcing applications. In this research paper, we introduce a novel Artificial Intelligence (AI) framework to support interactive and causality reasoning for a Chatbot application. The Chatbot will interact with the user to provide self-healthcare education and self-assessment (condition prediction). The framework is a combination of Natural Language Processing (NLP) and Knowledge Graph (KG) technologies with added causality and probability (uncertainty) properties to original Description Logic. This novel framework can generate causal knowledge probability neural networks to perform question answering and condition prediction tasks. The experimental results from a prototype showed strong positive feedback. The paper also identified remaining limitations and future research directions.Citation
Yu HQ (2020) 'Dynamic causality knowledge graph generation for supporting the chatbot healthcare system', Future Technologies Conference - Online, Springer Science and Business Media Deutschland GmbH.Additional Links
https://link.springer.com/chapter/10.1007%2F978-3-030-63092-8_3Type
Conference papers, meetings and proceedingsLanguage
enISBN
9783030630911ae974a485f413a2113503eed53cd6c53
10.1007/978-3-030-63092-8_3