The artificial intelligence continues to surprise us and it is that recently Goal new models of AI capable of recognizing and reproducing speech of over 1,100 languages, more than 10 times what technologies current ones can, in addition to achieving identify more than 4000 spoken languages.
Through a statement, the company ensured that their models Massively Multilingual Speech (MMS) expand the text-to-speech and speech-to-text technology from around 100 languages to more than 1, 100. Goal opened its new models and code so that others in the community of investigation they can develop your work and thus help to preserve the languages of the world and bring the world closer.
YOU MAY BE INTERESTED: Artificial intelligence will replace search and purchase sites, Bill Gates predicts
The company explained that, in order to collect data audio files from thousands of languages used religious texts, such as the Biblewhich have been translated into many languages different and whose translations have been extensively studied for the investigation of text-based language translation.
“These translations have publicly available audio recordings of people reading these texts in different languages. As part of the MMS project, we created a dataset of New Testament readings in more than 1,100 languages, which provided an average of 32 hours of data per language,” the statement explains.
ALSO READ: Learn about the 5 home devices that interfere with your WiFi
When considering unlabeled recordings from other christian religious readingsMeta managed to increase the number of languages available to more than 4 thousand . While this data is domain-specific and often read by male speakers, analysis shows that the models work equally well for male and female voices. And while the content of the audio recordings is religious, the analysis shows that this does not bias the model to produce more religious language.
This type of technology could be used for applications virtual reality and reality augmented in a person’s preferred language and who can understand the everyone’s voice
Follow us on GOOGLE NEWS for more news