- Date: 1-3 December 2020
- Location: Virtual
- Title: META Forum 2020
Iconic is pleased to announce our involvement in the second annual European Language Grid (ELG) conference, META-FORUM 2020, which takes place virtually 1-3 December. Our participation in the conference is through two of our European Commission funded projects, namely PRINCIPLE and MT4ALL.
The European Language Grid has been established as the primary platform for Language Technology in Europe. META-FORUM is an international conference series on innovative Language Technologies for the multilingual information society. Conference attendees will learn about the most recent developments in the European Language Technology industry and research, including language-centric AI.
Both the PRINCIPLE and MT4ALL projects are funded under the Connecting Europe Facility (CEF) programme. Both projects have a focus on low-resourced languages. The two-year PRINCIPLE Project, which started in September 2019, aims to provide language data to improve Machine Translation quality in the European Digital Service Infrastructures (DSIs) of eJustice and eProcurement. The under-resourced languages in this project include Croatian, Icelandic, Irish, and Norwegian – both Bokmål and Nynorsk. Iconic is pleased to have built bespoke Neural MT engines for use by a range of data providers from the four participating countries. These Early Adopters working with the PRINCIPLE project include: National University of Ireland Galway, Ciklopea, Icelandic Ministry of Foreign Affairs, Standards Norway and the Norwegian Ministry of Foreign Affairs. The next phase of the project will include validating the quality of project language resources through evaluation of MT systems.
The two-year MT4ALL project, which started in January 2020, is applying the latest advances in unsupervised MT to derive bilingual dictionaries and translation models, again for under-resourced languages, in the European Digital Service Infrastructure (DSI) areas of eHealth and eJustice. The project includes a selection of diverse languages aligned to targeted use-cases. The language pairs and use case domains are: Finnish, Norwegian, and Latvian to English in the financial domain, Ukrainian, Georgian and Kazakh to English in the legal domain, Norwegian, Spanish and German to English in the Customer Support domain, and Basque and Catalan to English in the General domain. Training of the first set of MT systems is currently in progress and the next phase of the project will involve a second round of system development and validations of MT systems.
Iconic is proud to be a part of these projects where cutting edge technologies are being further developed and deployed to enhance the Language Technology landscape for under-resourced languages. We are firmly committed to staying at the forefront of Language Technology innovation and to demonstrating leadership in putting advanced techniques into practice.
We would be happy to speak further about our Neural MT work at the META-FORUM. Please reach out and get in touch here if you would like more information.