Approximately 7,000 languages are spoken by people all over the world today. The remaining 3,000 languages are verbal and cannot be written down. Therefore, it has always been difficult to translate oral languages, and this frequently deepens the communication gap between people.
Now, Meta, the business formerly known as Facebook, has created the first-ever speech-to-speech translation system powered by AI that will predominantly interpret spoken languages.
According to Meta’s CEO Mark Zuckerberg, the business has created a speech-to-speech AI-supported translator that can translate Hokkien, a language that is largely spoken orally by about 49 million people in nations including China, Taiwan, Singapore, Malaysia, and the Philippines. The AI-backed translator is still in development and it has been open-sourced for researchers so that they can work on improving it.
Currently, Meta’s AI Hokkien translator can only translate one sentence at a time. Additionally, Meta provided a quick explanation of the new AI translator’s operation. The business has reportedly created a number of techniques for turning spoken words into a series of acoustic sounds that can then be used to generate waveforms using the ‘associated language.’ Hokkien uses Mandarin in this instance.
Meta asserts that it will provide free access to the technology and won’t restrict its speech-to-speech AI translation to Hokkien. The company’s translation technology is a component of its Universal Speech Translator project, according to a blog post on its official website.
Under this project, Meta is aiming to develop more AI methods to allow real-time speech-to-speech translation across many languages. With the new creation of AI translators, the company is also aiming to break the language barrier in both the physical and metaverse to establish better communication and understanding among people.
Post Your Comments