Der neue Brockhaus: Einsatz von sprachtechnologie und wissensnetz

Abstract

Human knowledge and language refer to and depend on each other. This fact is highly relevant if knowledge and language should be automatically processed in a way that allows the engineering of human knowledge. Outside of academic laboratory scenarios, scientific research in knowledge and language processing needs large, well-kept and properly encoded knowledge bases of language and domain knowledge. This challenge was successfully addressed by the company Brockhaus Duden Neue Medien GmbH in a research project together with the Fraunhofer IPSI institute in Darmstadt, the IICM institute of the Technical University Graz, the IAI institute of the University Saarbrücken and the KNOW-Center at Graz. For the new digital Brockhaus edition, special information representation and retrieval technologies were developed that are not restricted to one domain but cover the whole encyclopaedic knowledge space. The newly developed techniques were proven to be applicable to natural language search as well as to innovative knowledge visualisation. The results show - on the one hand - the opportunities for successful semantic retrieval without extensive application of formal meta-encoding. On the other hand, they point at the next steps to take in order to develop still more effective human-machine-interfaces.