That video is about 3 minutes too long, but the gist is that Meta is looking to advance its translation models to facilitate greater access, not just in the social media platforms of today, but more importantly, within the coming metaverse space.
As explained by Meta:
“Language is our lifeline to the world. But because high-quality translation tools don’t exist for hundreds of languages, billions of people today can’t access digital content or participate fully in conversations and communities online in their preferred or native languages. This is particularly an issue for hundreds of millions of people who speak the many languages of Africa and Asia.”
To improve this, Meta has been developing systems that are able to learn language translations from smaller datasets, while it’s also working with native speakers, where possible, to refine its systems.
Which has led to the development of its new translation model
“We’ve built a single AI model called NLLB-200, which translates 200 different languages with results far more accurate than what previous technology could accomplish. When comparing the quality of translations to previous AI research, NLLB-200 scored an average of 44% higher. For some African and Indian-based languages, NLLB-200’s translations were more than 70% more accurate.”
That will expand accessibility to more regions, while also ensuring that lesser used languages live on into the future, another important consideration.
But Meta’s systems alone won’t be able to facilitate full detection and translation of some languages. Which is why Meta is also open sourcing its data to invite more native speakers and experts into the development process.
“We’re also awarding up to $200,000 of grants for impactful uses of NLLB-200 to researchers and nonprofit organizations with initiatives focused on sustainability, food security, gender-based violence, education or other areas in support of the UN Sustainable Development Goals. Nonprofits interested in using NLLB-200 to translate two or more African languages, as well as researchers working in linguistics, machine translation and language technology, are invited to apply.”
In combination, these initiatives will help Meta evolve its translation tools, which, eventually, will enable users in the metaverse space to more easily converse and engage, in real-time, through language translation tools.
Google, too, is advancing its efforts on this front, with its translation tools now able to transcribe foreign language speech as it’s happening, providing more ways to interact in the moment.
The hope is that this will eventually facilitate more global connection and opportunity, by removing barriers from connection. But then again, that was the great hope of the internet, and social media as well – that, in providing a means to connect, we would facilitate more understanding and community, by enabling more people to join in the global conversation, and add more perspectives to enrich our understanding.
That’s not quite how things have played out, but there is a unique value to enhanced language translation, especially in regions where many languages are spoken, while it may also, eventually, facilitate Star Trek-like universal translator type tools within metaverse spaces that could open up whole new realms of connection and opportunity, in totally new ways.
Which is why this is an important project, and while it may be difficult to fully envisage as yet, it is good to see Meta looking to establish translation tools at the foundational level of the metaverse shift.
It could end up being a critical development – you can read more about Meta’s evolving translation process here.