SUMMA: Scalable Understanding of MultiMediA image

SUMMA: Scalable Understanding of MultiMediA

October 10th 2017 12:00 - 12:20

The SUMMA project ( brings together leading research institutions in the field of Natural Language Processing—the Universities of Edinburgh and Sheffield, University College London (all in the UK), Idiap Research Institute (Switzerland), Priberam Informática (Portugal), and the Qatar Computing Research Institute (Qatar) — with major national European news organisations—the BBC (UK), Deutsche Welle (Germany), and the Latvian News Organisation (LETA; Latvia) — to develop a new software platform for monitoring news streams, from TV news broadcasts to text to public social media feeds in multiple languages, thus allowing journalists to detect, track, and interpret developments in the world more quickly and more accurately, and also allowing multilingual news organisations to produce news more efficiently by fostering cooperation between news rooms across language barriers. Using cutting-edge Natural Language Processing Technology, including Deep Learning / Neural Networks, straight from the research lab, the platform will be able to automatically transcribe spoken language (e.g., from live broadcast or news video feeds on the internet) in several languages, translate transcripts and written text into English, recognise mentions of people, places, events, etc., determine how these entities relate to one another, and enter these reported facts into a factual knowledge base, facilitating researching background information, and an organisation's awareness of its own news reporting ("Have we reported on this elsewhere in our large, multilingual organisation?").

Peggy van der Kreeft avatar
Peggy van der Kreeft

Language Specialist - Project Manager at Deutsche Welle:

Language specialist with broad experience as a project manager in media and broadcast applications, documentalist, e-learning coach and evaluator at public examinations. Specialties: languages, media project management, ...


This website uses cookies. By continuing to browse you agree to this and Conferize's terms of service.