Nvidia launches a platform for creating interactive 3D avatars
At its annual GTC conference Tuesday, Nvidia announced Omniverse Avatar, a new platform for creating interactive AI avatars. The platform brings together Nvidia’s various technologies in speech AI, computer vision, natural language understanding, and simulation.
Nvidia says digital avatars created using the platform are interactive with ray-traced 3D graphics and can see, speak, and converse on various topics (via The Verge). The platform opens the door to the creation of virtual assistants that can be deployed in virtually any industry. Nvidia envisions the technology could help with a vast range of customer service interactions such as restaurant orders, making appointments and reservations, banking transactions, and so on.
During the presentation, Nvidia showed off a couple of demos of Omniverse Avatar. In one demo, we see a toy version of Nvidia CEO Jensen Huang answering questions on topics like biology and climate change. Meanwhile, another one shows an animated character in a restaurant kiosk conversing with customers as they order food from the menu.
“The dawn of intelligent virtual assistants has arrived. Omniverse Avatar combines NVIDIA’s foundational graphics, simulation and AI technologies to make some of the most complex real-time applications ever created. The use cases of collaborative robots and virtual assistants are incredible and far reaching,” said Jensen Huang, CEO of Nvidia.
Omniverse Avatar is part of Nvidia Omniverse, the company’s virtual world simulation, and collaboration platform. Initially limited to engineers and developers, Omniverse is now in Open Beta just about anyone and is currently being used by more than 70,000 users. BMW, Adobe, South Park, Bently, Ericsson, and Epigraph are among the companies that are using OmniVerse.
As mentioned above, Omniverse Avatar makes use of various speech AI, computer vision, and natural language processing technologies. For speech recognition and language understanding, it relies on NVIDIA Riva and Megatron 530B while NVIDIA Merlin and NVIDIA Metropolis provide recommendation engine and perception capabilities, respectively.