Takuya Nakamura, Philippe Capet, Thomas Delevallade, Julien Jacquelinet, Claude Martineau, Aude Rebotier, Agnes Sandor, Stavroula Voyatzi
Proceedings of the 3rd Language and Technology Conference. In Zygmunt Vetulani (ed.). October 5-7, 2007, Poznan, Poland, Widawnictwo Poznanskie Sp. z o o., pp. 539-543
ADAC is a risk assessment system endowed with intelligent awareness capabilities. It uses newswire texts (weak signals) as input, which are transformed into machine-processable data by human analysts. Owing to the large flow of information, human analysts cannot cope with this task any more. In this article we argue that this transformation can effectively be done by X-tract, a rule-based parsing tool for event extraction. In order to fulfill ADAC s requirements, besides extracting event mentions, X-tract also relates them to their information source and indicates their factuality. The integration of X-tract into ADAC is expected to imprve both its performance and its reliability
Report number: