Visualising Uncertainty and perspectives
The FP7 NewsReader project develops software to automatically ‘read’ massive streams of the news. News reports on changes in the world. Many different sources report on the same events. Partially, this information overlaps but partially different statements are made and/or different perspectives are expressed. In addition to a model to present the event information reported in news, we therefore also developed a model for the perspective of a source. This perspective reflects many different relations between the source and a statement, among which emotions, opinions, certainty, probability which are usually modelled separately. Using this model, we access information both at the level of statements but also at the level of the sources of statements: what are the statements of a source, how consistent is a source, what sources share the same statement, what source do not. The visualisation of these complex layers is a real challenge. The eScience project “Visualising Uncertainty and perspectives” will attempt to address this challenge. In this presentation, I will present the model and our first ideas about the visualisation.
Prof. Dr. P.Th.J.M. (Piek) Vossen (1960) is full Professor of Computational Lexicology at the VU University Amsterdam, Head of the Computational Lexicology & Terminology Lab (CLTL), co-founder and co-president of the Global WordNet Assocation (GWA) and partner in the Centre for Digital Humanities Amsterdam. In 2013 he won the prestigious Spinoza Award of the Netherlands Organisation for Scientific Research (NWO). He is also a member of the Koninklijke Hollandsche Maatschappij der Wetenschappen (KHMW).
Piek Vossen studied Dutch and General Linguistics at the University of Amsterdam and received his PhD (cum laude) in Linguistics on Computational Lexicology and Lexicography. He is/has been involved in the many national and international projects, among which: Acquilex, EuroWordNet, Meaning, Cornetto, DutchSemCor, KYOTO and GlobalWordNetGrid. His latest project NewsReader develops the ‘History Recorder’, a computer program that ‘reads’ the news each day and precisely records what happened when and where in the world and who was involved. Furthermore he developed 4 research projects with the Spinoza-price: “Understanding of Language by Machines – an escape from the world of language”.
For many years he combined his academic career with his work in the industry. He worked at Sail Labs (1999-2001) and as a C.T.O. of Irion Technologies B.V. (2001-2009), where he developed multilingual language technology for many different languages, such as cross-lingual semantic search, text classification, and natural-language dialogue systems.
His research interests are WordNets, Computational Lexicon, Ontologies, Computational Linguistics, Language Technology and Computer-Applications, both within a single language and from a multilingual perspective. Vossen is interested in the relation between lexicons and ontologies, from a theoretical point of view as well as from their usage in computer-applications in which meaning and interpretation play a role. He sees the lexicon as a fundamental resource to anchor meaning and interpretation in useful computer behaviour. Computer behaviour can make use of communicative models and insights from communication science. The organization of the lexicon and the knowledge stored in it need to take that usage as a starting point. He combines linguistics and computer science to model understanding of natural language texts by computers.