Simple Ontologies for Practical Information Extraction and Advanced Information Extraction for Practical Ontologies

Hong Li, Yi Zhang, Feiyu Xu, Hans Uszkoreit

In: Churen Huang, Yueguo Gu (editor). Contemporary Linguistics 当代语言学 2012/1 2012.


Information extraction can be regarded as a pragmatic approach to semantic understanding of natural language texts. Ontology is very important for modeling and specifying knowledge e.g. relations between entities and concepts. Therefore, ontology is often used for definition of the information extraction tasks. The advanced information extraction technologies such as complex relation extraction can be utilized for learning language patterns, which can recognize ontological relations from the free texts and extract relation instances. In this paper, we will describe an ontological model for information extraction tasks and present a general machine learning framework DARE for learning relation extraction patterns and extracting relation instances. DARE system has been intensively used for the English language. In this paper, we apply DARE to the Chinese newspaper texts and detect Chinese relation extraction rules and relation instances. Furthermore, we will compare our experiments with Chinese texts with the English texts.


OntologyIE-1.pdf (pdf, 2 MB )

German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz