Thread by @translation_eu

Good morning! We're back at #MetaForum2022. Day 2 is all about European Language Equality. @GeorgRehm now opening proceedings.

Live-stream:

The project is not about getting to full language equality. It's about setting up a strategy to get there, together with the many partners.

Today's first keynote comes to us from Luxembourg (moien!), courtesy of June Lowery-Kingston (@DigitalEU). June works on #accessibility, #multilingualism and safer internet.

@EU_Commission shapes policy, supports coordination and provides funding for inclusion and accessibility in the digital realm.

More details from @lk_june on funding opportunities

@lk_june All the useful multilingual resources from the European Commission can be found on this page (including multilingual tweets 😉): language-tools.ec.europa.eu

Challenges ahead for language equality:
- data availability and quality
- technical aspects (data processing/storage/formatting)
- legal aspects (privacy/GDPR/copyright)
- equality of support for all languages

Why #multilingualism matters: 50% of Europeans feel confident in English. We cannot exclude half the population! Conference on the future of Europe was a good example - @lk_june at #MetaForum2022

Next up: @F_Alfonsi MEP with his keynote on language equality.

After the contribution from Strasbourg, we now hear from the Basque country: German Rigau (@upvehu) investigates if language is represented in national lang-tech/AI strategies (alongside vision and robotics).

German Rigau works at @Hitz_zentroa, the Basque Center for Language Technology. The full report is available here: european-language-equality.eu/wp-content/upl… (PDF)

What is digital language equality anyway? @fedegasp1 shares his working definition: "all languages
have the technological support and situational context necessary for them to continue to exist and to prosper as living languages in the digital age". 👏🏼

@fedegasp1 Digital Language Equality can be measured and tracked over time using the DLE Metric:

@fedegasp1 Does this look equal to you? English the clear outlier, followed by German and English with roughly half the score of our lingua franca. This graph speaks volumes.

After the coffee break, session 8 brings us reports on the language support for Irish, Dutch, Bulgarian, Welsh and Icelandic. We're very curious!

First, a cross-language comparison with Maria Giagkou (@athenaRICinfo): EU official languages have better lang-tech support than other languages. Best support for English, Spanish, German, French. Least supported: Irish & Maltese. #MetaForum2022

Now, we look at the digital status of Irish with @cigilt: "fragmentary support". Not great, but an improvement over past years. Irish is also an EU working language, having left derogation status a while ago.

@cigilt Ireland's AI strategy mostly focuses on English language technology. There's currently no digital plan for #Gaeilge. Also: skills and funding gap, few Irish-speaking students in language technology.

Frida Steurs (@ivdnt) is up next to give us the latest on language technology for Dutch. Things actually looking good - e.g. Dutch present in most commercial translation services (Google, Bing, DeepL) and EU's #eTranslation

How about български? Svetla Koeva is back with an update: #MT still underdeveloped and many technologies unavailable, but number of language models growing, improving ICT education in schools, many AI companies.

Gareth Watkins with the latest on Welsh language technology support:
- Government Action Plan
- Collab w/ @Mozilla on speech data gathering
- new Welsh synthetic voices
- Lleisiwr project with @NHSuk for personalized bilingual synthetic voices
- Trawsgrifiwr (transcriber) etc.

#Icelandic lang-tech support is a political priority: President Jóhannesson went to California to talk directly to tech companies and support for his country's language. icelandmonitor.mbl.is/news/culture_a…

@eirikurr has more details on new resources and tools, such as corpora, #xl8 models, text-to-speech/speech-to-text and #MT for Icelandic. #MetaForum2022

icelandmonitor.mbl.is/news/culture_a…

thejournal.ie/minority-langu…

Session 9 is underway now: a closer look at the European lang-tech community. Input from @CLARINERIC and @vision_claire. #MetaForum2022

Some very interesting survey results from @_NatResende on European lang-tech users and citizens in general. What do they use? How much do they know about the technology? So much to take in, we'll need to get our hands on the slides.

Moderated by @cigilt, we now dive deep into 4 different sub-fields:
- Data, Language Resources, Knowledge Graphs
- Text Analytics, Text and Data Mining, NLU (Natural Language Understanding)
- Machine Translation
- Speech Technologies
Where are the gaps and shortcomings?

Language technology absolutely needs big amounts of quality data. #GDPR has made this more difficult, especially in domains like health. Same for legal aspects (#licensing and #copyright) with regard to translation of fiction, for example. (@AndrejsVas)

As we move towards the end of #MetaForum2022, we look at the future of the ELE and ELG projects - the Strategic Research and Innovation Agenda (SRIA). What needs to happen next to make meaningful progress? For example in terms of data availability, data bias or "green lang-tech".

Two interesting ideas, presented by @GeorgRehm:

1) make multilingualism and lang-tech part of a Commissioner portfolio (again)

2) require all EU projects to include a language diversity plan

👏🏼👏🏼👏🏼

Both ELE and ELG have accomplished remarkable things. But a lot still needs to be done. What exactly? @GeorgRehm, @JaneDunne5 & Jan Hajic tell us more at #MetaForum2022.

@GeorgRehm @JaneDunne5 @GeorgRehm with a call to "unlock the treasure chests" containing valuable data around Europe! #MetaForum2022

Good news: Two books to be published very soon (ELE/ELG) and META-Forum will most likely return in one year's time. 🥳

And that's a wrap! A big thank you to the organising team on stage and behind the scenes and to all speakers and presenters! See you at the next META-Forum.

Some more interesting reading for #metaforum2022 attendees and followers:

https://twitter.com/emilymbender/status/1534525537804161024

https://twitter.com/adrechsel/status/1534169160921513986