We're pleased to announce that you can try Bergamot's local translation technology in your browser today, with the Firefox Translations extension: https://t.co/l09hvAivHv
Please try it out and take the in-product survey to provide us with feedback!
Visit our virtual booth at #MetaForum2021 11:00-12:30 GMT for a local privacy-preserving translation demo including one you can run yourself. Registration is free: https://t.co/rsg0ZRgdXX . #HorizonEU
🏴 @EdinburghUni's Kenneth Heafield joins SlatorPod 🎙️ to discuss #translation efficiency, quality research, and industry vs. academia. Check out the super fast translation app here 👉 https://t.co/JSg5nEKZ5o
@zngu #t9n#xl8#NLP#MT#data#tech
https://t.co/5P02IqUe7K
Sneak peak of some incredible progress on integrating Neural Machine Translation in a browser extension. Check out a video of the Bergamot prototype WebExtension for #Firefox https://t.co/xmiCCK4Mfb
Two more exciting open roles to help advance the state of Neural Machine Translation (NMT) models, collaborating with the Bergamot team:
https://t.co/Glrcnj0i60
and
https://t.co/jVbtgU7hxU
Tell your friends :)
Want to work on Developing the future of private, client-side machine translation?!?
The Bergamot project is hiring for a postdoctoral research associate Closing date 9th October: https://t.co/Qjtd5esrTi
Any native English speakers interested in helping the @BergamotProject generate user data to calibrate the quality of #private local #machinetranslation ?
https://t.co/2BZNWaeBni
Great presentation by @NikolayBogoychev at this year's W3C Workshop on Web and Machine Learning showing the recent progress of Bergamot's privacy focused machine translation running locally in Firefox!
https://t.co/k4INQikWMo
@marian_nmt The Google and Microsoft translation mobile apps do offline, that's a lot of devices.
Interestingly I see a lot of requests for Google Translate on-prem. It's possible to hack it from the offline language packs, but will be much easier to get it from open-source Firefox.
@ArgoTranslation We train models in advance using @ParaCrawl, https://t.co/gDU57w6Buh, and other data. Then these models are downloaded to the user's computer for inference. Our domain adaptation work package led by Tartu aims to tailor models to users and websites locally.
The Bergamot project is a consortium of the University of Edinburgh, Charles University in Prague, the University of Sheffield, the University of Tartu, and Mozilla.
Efficient browser translation by @marian_nmt, quality estimation, and help filling forms in other languages.