NVIDIA announced the Omniverse Avatar platform, which will enable the creation of AI-based 3D avatars ready for the Metaverse
At the GTC event NVIDIA announced Omniverse Avatar, a platform that enables the generation of AI-based interactive 3D avatars. A definition that contains a strong reference to the Metaverse announced by Mark Zuckerberg a few days ago, which will need billions of avatars to welcome users.
The strong point of the Omniverse Avatar platform is that it manages to exploit and bring together the best graphics technologies developed over the years by NVIDIA in order to produce high-level 3D avatars. The “characters”, thanks to artificial intelligence, will be interactive and intelligent, able to perceive what is around them and to hold conversations on a wide range of topics. “Omniverse Avatar,” the company explains, “directs the company’s technologies on voice AI, computer vision, natural language understanding, recommendation engines and simulation technologies.”
NVIDIA’s Omniverse Avatar Goals
The Omniverse Avatar 3D avatar platform opens the door to creating AI-based assistants that are easily scalable to any industry. Avatars, in addition to being able to come in handy for the Metaverse of Meta, could provide valuable support for the billions of daily interactions borne by customer services across a wide range of industries, from restaurant orders and banking transactions to appointments and reservations for goods or services, and more, “leading,” NVIDIA says, “to more business opportunities and greater customer satisfaction.”
Jensen Huang, founder and CEO of NVIDIA, expressed satisfaction, saying, “The dawn of intelligent virtual assistants has arrived. Omniverse Avatar combines NVIDIA’s core graphics, simulation and AI technologies to build some of the most complex real-time applications ever created. Collaborative robots and virtual assistants can find use in an incredible number of scenarios.”
It’s no coincidence that Huang, the CEO, mentioned virtual assistants. NVIDIA’s Omniverse Avatar aims, among others, to contribute to the creation of a new generation of virtual assistants that are even more AI-driven than the current Amazon’s Alexa, Google’s Assistant or Apple’s Siri. The time, according to NVIDIA, would be ripe for a generational shift that offers the public “truly” interactive assistants, 360 degrees, perhaps to be found then within the Metaverse.
NVIDIA technologies for Omniverse Avatar
The technologies employed by NVIDIA for Omniverse Avatar are multiple. Speech recognition is based on NVIDIA Riva, a software development kit that recognizes speech in multiple languages; language understanding is provided by the Megatron 530B pretrained model that can recognize, understand and generate human speech; and the NVIDIA Merlin framework enables companies to create deep recognition systems that can handle large amounts of data to provide smarter suggestions.
This is, therefore, a huge gamble for NVIDIA and the latest major investment by a chip big in the Internet of the future, coming just hours after Meta and AMD closed their deal.