Multilingual Event Extraction for Border Security Intelligence Gathering
Jakub Piskorski

Joint Research center
European Commission


This talk gives an overview of an effort on deploying news event extraction technology for border security intelligence gathering and real-time situation monitoring for Frontex, the European Agency for the Management of Operational Cooperation at the External Borders of the Member Stated of the European Union. In particular, a hybrid multilingual event extraction system has been constructed on top of the Europe Media Monitor, a large-scale news monitoring aggregation engine. The hybrid system integrates two existing event extraction engines, namely, NEXUS - developed by the Joint Research Centre of the European Commission, and PULS - developed by the University of Helsinki. The presentation explains the entire event extraction processing chain and highlights various aspects of information access, moderation and visualization.

Short bio

Jakub Piskorski received his M.Sc in Computer Science from the University of Saarbrücken, Germany in 1994 and Ph.D from the Polish Academy of Sciences in Warsaw, Poland in 2002. His areas of interest are centered around finite-state technology, shallow text processing, information extraction, efficient application oriented natural language processing solutions and open source intelligence. Jakub is currently working in the Research & Development Unit of the Warsaw-based EU Border Security Agency Frontex and he is also a Research Associate at the Polish Academy of Sciences in Warsaw. Previously he has worked for the Joint Research Centre of the European Commission, the German Research Centre for Artificial Intelligence in Saarbruecken and the Department of Information Systems at Poznan University of Economics. He also has been consulting several companies on text mining and information extraction technology. Jakub is author and co-author of around 80 peer-reviewed international conference papers and journal articles, which cover various topics related to natural language processing, text mining and security applications.