GoURMET, an H2020 project focused on (better) neural machine translation (NMT) for low resource languages and domains, is about to cross the finish line. With the final review only weeks away, it’s time to take a look at what the consortium created over a course of more than three years (January 2019 – September 2022).
(more…)
Great news from the world of HLT (that’s Human Language Technology): In the scope of the GoURMET project, our partners at the BBC have completed a very useful tool called GoURMET translate. It helps non-expert users translate text from and into a number of low-resource languages – currently Amharic, Bulgarian, Gujarati, Kyrgyz, Serbian, Swahili and Turkish. (more…)
If you’ve been following this blog, you’ve probably heard of SUMMA, one of our biggest and most succesful language technology projects so far. Over the course of three years, the SUMMA consortium (= DW and 8 other partners) has built a comprehensive platform for multilingual media monitoring. But what does that mean exactly? And how will the software benefit newsrooms? Our explainer video can answer those questions in less than two minutes:
(more…)
SUMMA is an ambitious research and innovation project aimed at media monitoring, with a close collaboration between Deutsche Welle and the BBC. Through language technologies such as speech-to-text (automated transcription), machine translation, topic detection and summarization, the SUMMA platform will help media professionals, amongst others, to monitor and analyze news more efficiently.
The project will capture incoming news streams in eight languages, analyze and compare the data, and provide an overview of leading topics and trends with summaries in English. The University of Edinburgh is leading the project, consisting of eight partners.
The consortium is aiming at jointly developing three tools :
Deutsche Welle will be working towards testing and evaluating the entire SUMMA platform, and is the primary user partner for the internal monitoring tool. It is envisaged that the work will lead to a user-friendly and tested tool, specifically tailored to analyzing DW content, with potential for subsequent integration and implementation.
SUMMA runs from February 2016 – January 2019.
Bring me to the SUMMA website
Metadata and Machine Translation
When: 11 June 2015
Where: EBU Headquarters, Geneva, Switzerland
What is it?
The European Broadcaster Union’s SCAIE Workshop (Study of Content Analysis-based Automatic Information Extraction in Production) on “From Research to Production” wants to bring together experts in automatic metadata extraction, ICT specialists and broadcast industry stakeholders involved with bringing results of research on automatic metadata extraction into production workflows.
(more…)