大规模知识图谱管理与挖掘【精品】.pdf

Managing and Mining Knowledge Graphs – Challenges and Opportunies 肖仰华 复旦大学 GDM@FUDAN hp:// Copyrights •  Fabian Suchanek Gerhard Weikum, Knowledge Harvesng in the Big Data Era, SIGMOD 2013 Tutorials. •  Bin Shao, Haixun Wang, Yanghua Xiao, Managing and Mining Large Graphs: Systems and Implementaons, SIGMOD 2012 Tutorials. •  肖仰华，面向知识图谱的数据融合与管理，NSFC重点基 金“网络数据融合理论与技术”研讨会，苏州， 2013-8-14. •  肖仰华，中文知识图谱， 上海， Italk 沙龙，2012 Outline •  Preliminaries •  Opportunies •  Challenges •  Managing big knowledge graph •  Mining big knowledge graph What is knowledge graph? knowledge graph contains enes/concepts as verces and semanc relaonships as edges Chinese Knowledge Graph 4 What makes knowledge graph different ？ •  Ontology •  Knowledge graph –  Domain dependent –  Large scale –  Small scale –  Cover both enes and –  Edited by humans concepts –  Cover different semanc relaonships •  Semanc network –  Automacally harvested –  Focus on concepts from Web or other large instead of enes scale corpus –  Low coverage Chinese Knowledge Graph 5 Web of Data: RDF, Tables, Microdata 60 Bio. SPO triples (RDF) and growing