Meta announces launchAn artificial intelligence model called SeamlessM4TIt will be able to recognize nearly 100 languages, as well as provide text-to-text and voice-to-text translation functions. In addition, it can convert nearly 100 language sounds into 36 different languages including English, or output them into 35 languages including English.
SeamlessM4T transmits approximately 27 hours of voice and text contentConduct trainingAt the same time, compared with other language translation tools, it only covers some of the world's language content. Therefore, when translating between languages with different cultural backgrounds, there are often obvious errors, or the details are not detailed, and even the entire translation efficiency is low. Meta said that SeamlessM4T combines multiple language translation models to achieve more efficient and correct translation between multiple languages.
Meta CEO Mark Zuckberg believes the launch of SeamlessM4T will facilitate communication between people speaking different languages, especially in virtual worlds like the Metaverse, where direct language communication remains a crucial medium of interaction. Furthermore, Meta will make the SeamlessM4T model publicly available for non-commercial use.
In addition to Meta, companies including Google, Microsoft, and Amazon continue to update their translation tools. OpenAI and other companies are also using artificial intelligence technology to enhance natural language understanding and translation functions, allowing users to interact with artificial intelligence through natural language.
