OPUS-CAT MT : à la découverte d’un logiciel de traduction automatique neuronale installé en local
#TraductionAutomatique #Multilinguisme #ScienceOuverte #Bibliodiversity #OpenScience #MachineTranslation par Jean Francois Nominé INIST-CNRS
https://lalist.inist.fr/?p=61931

Err... No thanks, DeepL. This word does not even exist.

Be careful with MT!

#xl8 #l10n #machinetranslation

I would love to opt-out of #MachineTranslation for stuff I write in #Icelandic. It simply doesn't work. For larger languages it makes frequent mistakes. For smaller languages it can garble the meaning completely. It has no sense of nuance, subtlety or humor.

It is alarming how many people think they can just #AutoTranslate anything.

Meta launches an AI model / SeamlessM4T / which can transcribe and translate almost 100 languages.

https://seamless.metademolab.com/demo

#Meta #MetaAI #SeamlessM4T #SpeechToText #Translation #AIpoweredTranslation #UniversalSpeechTranslator #MultilingualTech #OpenSourceAI #SpeechRecognition #LanguageBarrierBreaker #AIinCommunication #MetaResearch #MachineTranslation  #NextGenTranslation #AI

Gizmodo: Meta Releases AI to Translate Dozens of Languages Using Speech and Text https://gizmodo.com/meta-releases-seamlessm4t-translation-ai-text-speech-1850761868 #applicationsofartificialintelligence #massivelymultilingualspeechsystem #computationallinguistics #massivelymultilingual #technologyinternet #machinetranslation #largelanguagemodel #speechrecognition #multilingualism #speechsynthesis #creativecommons #humaninterest #translation #pacoguzman #metaai #meta #paco

If someone wants the same job as I have: Our group at the #UniversityofEdinburgh of is looking for researcher assistants/associates with strong technical skills to work on #nlp projects! If you'd like to dabble with 12PB of web data, or #LLMs, or #machinetranslation and related things: https://elxw.fa.em3.oraclecloud.com/hcmUI/CandidateExperience/en/sites/CX_1001/job/8196

I didn't have high hopes from the slightly clickbaity title, but found this quite a reasonable, well balanced article.

"To polyglots, foreign languages are Mount Everests daring us to climb them. But to most people, they are just a barrier to get to the other side of."

"Learning to genuinely speak a new language will hardly be unknown. It will continue to beckon, for instance, for those actually relocating to a new country. And it will persist with people who want to engage with literature or media in the original language, as well as those of us who find pleasure in mastering these new codes just because they are “there.” In other words, it will likely become an artisanal pursuit, of interest to a much smaller but more committed set of enthusiasts."

#Translation #MachineTranslation #Language #LanguageLearning

https://www.nytimes.com/2023/07/25/opinion/translation-apps-foreign-languages.html?unlocked_article_code=ZPiaG-vh0nKT6EIPtzXl7g94GitTI9HLjTmB5lULwyv6cq5SPrkcr-WtXPKb6t_O57sBAFolYN-3b5qQHGTTNyY1whoHBtteywUeMLwcTUyi5VbiHBUCLqSzUOOdAWlp9kKrgloQDsXLPhLf_7g4OfMDKtgMdI3XCxn7heTE6GDo8Ql7pjXiGIi1pnvz5sGaSYWqkcYRJRq-eKi11S0tat4988LalEKfplz-XPaNME0HLNZvwKwoi5yHYYsvoWCh_OkZlYKG-u0bRzrNlnxcdxP3ga0cftwed_PU_42XF8_tpQYQ-5wjjGvR5iGD-fTlzGu5RKPC3PGT4wBTJxNioJ1oHm7A2zaNlydcaQ&smid=url-share

Need to go from “male X” to “female X” or vice versa? Here’s a handy tutorial on how to do gender rewriting with the Fairslator API: https://rapidapi.com/lexiconista/api/fairslator/tutorials/how-to-rewrite-genders-with-the-fairslator-api #MachineTranslation #GenderBias

I can’t stress enough how *simple* gender rewriting is as a technique for dealing with #GenderBias in #MachineTranslation. Here I’ve summarized it in three easy steps. https://www.fairslator.com/how-gender-rewriting-works

Scholars in #MachineTranslation might be interested in this repository to help you #evaluate your model with various #metrics:

https://github.com/h9-tect/Machine_Translation_Evaluation_Metrics

Summaries of pre-prints on #MachineTranslation and multilingual #NLProc that I liked the most in June are on my blog 🤓 https://jlibovicky.github.io/2023/07/08/MTML-Highlights-June.html

MinT: Supporting underserved languages with open machine translation

https://diff.wikimedia.org/2023/06/13/mint-supporting-underserved-languages-with-open-machine-translation/

"MinT ('Machine in Translation') is a new translation service by the Wikimedia Foundation Language Team that is based on open source neural machine translation models and supports over 200 languages. ...The translation models used by MinT support over 200 languages, including many underserved languages that are getting machine translation for the first time."

#MachineTranslation #OpenSource

The paper and pre-prints that I found interesting in May are on my blog now https://jlibovicky.github.io/2023/06/08/MTML-Highlights-May.html #NLProc #MachineTranslation #Multilinguality

Are there any existing findings regarding the quality of GPT translation in real-world situations? Has it been proved to be better than Google Translate, Microsoft Translator or any other commercial MT engine? #machinetranslation

I just discovered that if you put Korean text into Google Translate and set the source language to Japanese, sometimes it will output something completely wrong ("hello", "I'm sorry", etc). When the source language is set to Chinese it just spits the source text back out. This could be consequential in cases where language auto-detection is wrong.
#MachineTranslation #GoogleTranslate

Teaching data literacy in the context of MT literacy 💻 @th_koeln guided through an example of their learning resources and showed how they are structured and how they can be used:
#machinetranslation #translation #xl8
@J_Hackenbuchner @manuela_schloe https://youtu.be/-lZZ1tcAATQ

🐦🔗: https://n.respublicae.eu/translatores/status/1654002951680327680

🌍 Machine Translate's API pages have just been updated!
Numerous MT APIs have recently included support for new languages, like ModernMT reaching 200!
Check it out ➡️ https://machinetranslate.org/apis

#machinetranslation #machinetranslate

#ChatGPT has poured fuel into the ongoing discussion on whether #machinetranslation can keep up with human translation, but fortunately, brighter flames shed more light on the topic.
Jost Zetzsche has written a remarkable article on what #MT (or any text-processing #AI) is good for and when (not) to use it on the ATA Blog: https://www.atanet.org/client-assistance/machine-translation-vs-human-translation/

AI is revolutionizing bad work in many fields. (via @RespondCrisis on twitter)

#artificialintelligence
#machinetranslation
#AIscam

https://restofworld.org/2023/ai-translation-errors-afghan-refugees-asylum/

5 real-world applications of natural language processing (NLP) - Chatbots, sentiment analysis, speech recognition, text summarizat... - https://cointelegraph.com/news/5-real-world-applications-of-natural-language-processing-nlp #naturallanguageprocessing #real-worldapplications #machinetranslation #sentimentanalysis #textsummarization #speechrecognition #chatbots #nlp