報告人：陳矯彥 博士 曼徹斯特大學
主持人：漆桂林
報告時間：2025年1月10日（周五）下午2:00
報告地點：bet356手机版唯一官网九龍湖校區計算機樓313室
報告摘要：Ontology plays an important role in information systems and AI in many domains like e-commerce and biomedicine, due to its capabilities for knowledge representation and semantic reasoning. The construction of reliable ontologies, however, still relies on experts, costing a huge amount of human labour. With the success of Large Language Models (LLMs), more and more efforts are to automate ontology construction by combining LLMs with knowledge representation and machine learning techniques. In this report, I will introduce our library DeepOnto which includes a series of APIs for utilising LLMs for ontology construction tasks such as ontology alignment and completion, as well as recent research combing different LLMs for ontology completion. As a fundamental technique for ontology construction and neural-symbolic AI, ontology embedding, which is to represent complex knowledge in a vector space, has become a promising direction with wide attention. Thus, I will also briefly introduce our recent progress on ontology embedding either using LLMs or not.報告人簡介：Dr. Jiaoyan Chen is a Lecturer (Assistant Professor, tenured) at the Department of Computer Science, The University of Manchester, and works as a part-time Senior Researcher for the Department of Computer Science, University of Oxford. He mainly studies Knowledge Graph, Ontology, Knowledge Representation and Machine Learning, with a recent focus on the integration of Knowledge Graph and Ontology with Large Language Models. Jiaoyan has served as PC, Senior PC and reviewers for many top Computer Science conferences and journals such as ISWC, WWW, IJCAI, AAAI and Journal of Web Semantics, and is now serving a new journal named Transactions on Graph Data and Knowledge as Associate Editor. Home page: https://chenjiaoyan.github.io/
