None
Brief news summary
Facebook and Instagram will soon be more accessible in lesser-spoken languages, but experts suggest that Meta, the parent company, should engage with native speakers to improve the translations. Meta's No Language Left Behind (NLLB) project has expanded its technology to include 200 languages, including Scottish Gaelic, Galician, and Welsh. Meta trains its AI using data from the Opus repository, and the translations are evaluated using benchmark sentences. However, experts argue that talking to native speakers and linguists is vital for enhancing the translations. They also suggest using high-quality data sources and complying with legal requirements for better results. Despite the progress, some experts believe that the translations are not yet reliable, while others emphasize the importance of using the tools to motivate further improvements.The No Language Left Behind (NLLB) project, led by Meta, aims to make Facebook and Instagram posts more accessible in 200 lesser-spoken languages worldwide. The project has successfully scaled its original technology, including languages like Scottish Gaelic and Bosnian. However, experts suggest that Meta should consult with native speakers and language specialists to further improve the tool. Meta trains its AI using data from the Opus repository and combines it with mined data from sources like Wikipedia.
The accuracy of translations has improved by 44% since the first model was released in 2020. Despite the progress made, some experts question the accuracy and quality of translations, recommending that Meta collaborate with organizations like the BBC for better language preservation. Linguists and technical specialists agree that refining data sets and working together is crucial. While some believe users should give Meta translations a chance to encourage improvement, others caution against relying on the tools due to potential errors.
Watch video about
None
Try our premium solution and start getting clients — at no cost to you
