This Wednesday, February 23, 2022, Meta organized a virtual event called Inside the Lab. For nearly two hours, the firm unveiled its new advances in the field of artificial intelligence, which it considers ” one of the keys to the metaverse “.
The development of the metaverse goes through AI
” The kind of experiences you will have in the Metaverse are beyond anything possible today. It’s an immersive version of the internet: instead of looking at something on a screen, you feel like you’re inside or present with another person. This is going to require advances in a whole host of areas, from computing devices to software for building and exploring worlds. “, thus declared Mark Zuckerberg.
AI: fake faces that are increasingly difficult to identify
According to the CEO of Meta, artificial intelligence is an essential component in the development of the metaverse, on which the company is betting its entire future. Its specialized AI division is therefore currently working on new technologies that will make it possible to build the metaverse. ” Much of AI research today is focused on understanding the physical world. But in the metaverse, we will need an AI designed to help people navigate virtual worlds as well as our physical world with augmented reality. And since these worlds will be dynamic and constantly changing, the AI will need to be able to understand the context and learn like humans do. Zuckerberg continued.
The firm thus took advantage of Inside the Lab to present some of its advances in this area.
Break the language barrier
First of all, it looks like Meta literally wants to break down the language barrier in the metaverse so that everyone can instantly communicate with everyone. To achieve this, it therefore presented an initiative that includes two very distinct projects.
The first, No Language Left Behind, ” is to develop a new advanced AI model that will be able to learn languages without needing as many training examples as current models “explains Meta. The second, Universal Speech Translator, “ aims to devise new approaches to translating spoken language from one language to another in real time, in order to support languages that do not have a standardized writing system as well as those that are both written and spoken “.
Meta says it still faces challenges to achieve such results, particularly with respect to the lack of data needed to train the models. ” The effectiveness of these and other technologies will increase when they can work in the same way for billions of people around the world, giving them equal access to information and enabling them to communicate with a much wider audience, whatever languages they speak or write says the company.
A revolutionary voice assistant in the metaverse
Mark Zuckerberg’s firm also announced the development of the CAIRaoke project, which aims to revolutionize the way humans communicate with their devices. She thus created a neural model of AI based on machine learning, that is, a statistical model based on the biological networks of the human brain. The latter uses the self-supervised learning meaning that instead of being trained on large data sets like many other models are, this AI can essentially learn on its own.
She wants to incorporate this technology into a voice assistant that could accompany users within the metaverse, and would therefore be dedicated to augmented reality and virtual reality. Mark Zuckerberg also demonstrated the amazing capabilities of the model using an AI called Builder Bot, which allows you to generate or import elements into a virtual world by simple voice command.
” On devices such as virtual reality headsets and augmented reality glasses, we expect this type of communication to become the primary method of navigation and interaction, just as touchscreens have replaced keyboards on smartphones. says Meta.
Zuckerberg promises transparency
The period that Meta is going through is complex. After its name change at the end of 2021, the firm demonstrated that it wanted to put past scandals behind it to focus on its future and on the development of the metaverse. In this context, Mark Zuckerberg wanted to be reassuring by promising full transparency on the progress of his company.
” We are committed to building openly and responsibly. By moving forward in this direction, we have the opportunity to create better and safer online environments for all of us. This means creating AI technologies that provide the highest levels of privacy and help prevent problems […] This means engaging with experts in human rights, civil rights, disability rights and privacy and building systems based on fairness, respect and human dignity. This means working in an open way, and sharing progress. And that means building the metaverse for everyone, so people around the world have access to the tools and technologies to realize their own vision for the future. “said the CEO.
Meta’s various advances in AI show that the firm is working hard to deploy the metaverse. While it still has a lot of work and technical challenges left, these various announcements allow us to get a glimpse of what the future iteration of the Internet will look like.