Domain-Adaptive Relation Extraction for the Semantic Web

Abstract

In the THESEUS Alexandria use case, information extraction (IE) has been intensively applied to extract facts automatically from unstructured documents, such as Wikipedia and online news, in order to construct ontology-based knowledge databases for advanced information access. In addition, IE is also utilized for analyzing natural language queries for the Alexandria question answering system. The DARE system, a minimally supervised machine learning system for relation extraction, developed at the DFKI LT-Lab, has been adapted and extended to the IE tasks for Alexandria. DARE is domain-adaptive and has been used to learn relation extraction rules automatically for the Alexandria-relevant relations and events. Furthermore, DARE is also applied to the Alexandria opinion mining task for detecting opinion sources, targets and polarities in online news. The DARE system and its learned rules have been integrated into the Alexandria IE pipeline. © Springer International Publishing Switzerland 2014.