Information Extraction in a Multilingual Context
Maarten de Rijke, Informatics Institute, University of Amsterdam, The Netherlands
Abstract
With the increasing focus on retrieval systems that return information rather than just documents, robust and scalable information extraction (IE) technologies quickly gain in relevance. Opinions, themes, stakeholders, entities as well as their relations and attributes need to be identified reliably. In a European setting, and with tasks such as news monitoring, open source intelligence, and social media analysis, information extraction gets a distinctly multilingual flavor. In the presentation, some of the tasks giving rise to the challenges in multilingual IE will be discussed, together with proposed solutions. The presentation will be organized around examples and solutions that complement the ones offered in Anselmo Penas's presentation.
Course Material
Information Extraction in a Multilingual Context
Maarten de Rijke, Informatics Institute, University of Amsterdam, The Netherlands
