Neural-symbolic Knowledge Representation with Ontology and Knowledge Graph Embeddings

Overview

Abstract

Ontologies and Knowledge Graphs are becoming increasingly popular for knowledge representation and reasoning, with a fundamental role in AI and Information Systems. Their embeddings are to represent entities in a vector space with their formal and informal semantics (such as logical relationships and textual meta information) concerned. In this talk, I will first briefly introduce the definitions of ontology, knowledge graph and their embeddings, then introduce some embedding methods, including those using geometric modelling and (large) language models, and finally discuss the role of Knowledge Graph, Ontology and their embeddings in semantic reasoning such as knowledge completion, and in addressing Machine Learning challenges such as sample shortage and augmenting large language models.

Brief Biography

Dr. Jiaoyan Chen (https://chenjiaoyan.github.io) is a Lecturer in Department of Computer Science, The University of Manchester, and a part-time senior researcher in Department of Computer Science, University of Oxford. His research interests mainly lie in Knowledge Representation, Knowledge Graph, Ontology, and Machine Learning, with over 10 years research experience and over 60 papers published in top Computer Science conferences and journals including AAAI, IJCAI, WWW, ISWC, ICLR, PIEEE, Machine Learning, JWS and so on. Jiaoyan is also an Associate Editor of Transactions on Graph Data and Knowledge. In 2023, Jiaoyan was awarded a New Investigator project OntoEm by EPSRC for supporting his research on neural-symbolic knowledge representation.