Today’s Invention was Yesterday’s Science Fiction. The “oddest thing in the cosmos”, the Babel Fish, is a type of fish depicted in Douglas Adam’s magnum opus, The Hitchhiker’s Guide to the Galaxy. The fish worn as an earpiece quickly translates that have ever been.
Babel Fish is no longer a pipe dream because of improvements in AI, particularly in the NLP arena, numerous tech titans are working on a universal translation. To that purpose, Meta focused on the Universal Speech Translator in its Inside Lab event on Feb 23rd,2022.
Billions of people are still unable to access most internet content or engage readily with the majority of the online world in their native language. Machine Translation (MT) systems are quickly improving, but they still rely significantly on learning from enormous volumes of textual input, making them unsuitable for low-resource languages.
Eliminating communication barriers would have far-reaching consequences, allowing billions of people to access information online in their native or chosen language. Advances in MT will not only benefit those who do not understand one of the
languages that now dominate the internet; they will also profoundly revolutionize the way people across the world communicate and exchange ideas.
According to Facebook, around 25% of the world’s population speak a language without a translation system. The firm is working on two initiatives to address this issue. One section is titled “No language Left Behind”, and the second is titled “Universal Speech Translator”.
On Feb 23rd, during the Inside the Lab, building for the Metaverse with AI webcast event, CEO Mark Zuckerberg provided a preview of what Metaverse will look like. He discussed AI’s most recent achievements in self-supervised learning and computer vision. Mark explained that the company’s goal is to create an AI-powered “Global Voice Translator” for the Metaverse that is accessible to everyone on the planet.
The AI developer will be working on the projects which will provide:
“No Languages Left Behind”
The team will create a sophisticated AI model that learns languages from only a few samples. The model is capable of producing expert-quality translations in hundreds of languages.
“Universal Speech Translator”
The technology will immediately translate speech in real-time, eliminating the need for text conversion. Mark sees this as a chance to expand on its already popular concepts. The translation programme would improve the capabilities of the proposed AR Glasses while also remaining constraints in Metaverse’s VR and AR reality environments. Meta’s concept would be widely accepted because of the widespread use of AI development services, adopting it as translating tools.
It is undeniable that improvements in machine learning in recent years have greatly increased the speed and accuracy of machine translation. Several major technology firms, such as Google to Apple, are now providing free artificial intelligence services on translation tools for business and tourism, which will assist people worldwide.
However, the technology is reported to have its own set of issues. It is predicted to generate problems such as gender bias, missing subtitles that are important for human speakers, and possible other ludicrous mistakes. Many people who speak one of these rare languages are afraid of losing their culture.
Mark added that they developed the first multilingual text translation system that is not English-centric to scale text-based MT to 101 languages. Typically, bilingual systems function by translating from the source language into English and then into the target language. To improve the efficiency and quality of these systems, they removed English as a medium, allowing languages to be translated directly into other languages without going via English.
The firm wants its technology to be as inclusive as possible; thus, it should accommodate written languages that do not have a conventional writing system. In light of this, the team is working on a speech-to-speech translation system that does not rely on producing an intermediate textual representation during interference.
Hopping to the Future of Speech
Mark said, if in collaboration with the AI solutions and MT research community, No Language Left Behind and Universal Speech Translator successfully develop translation technologies that encompass everyone globally; it will open up the digital and physical worlds in unimaginable ways.
They’re already making progress in allowing translations for low-resource languages, which are a substantial obstacle to universal translation for the vast majority of the world’s populations. Aiming that by developing and open-sourcing the work, multilingual modelling, and assessment, other academics will build on it and bring real-world translation system applications closer to reality.
