A huge crowd of more than 2,600 people turned out for NVIDIA’s fifth regional GPU Technology Conference, held here in Tokyo today. This follows sellout GTC events in the last few weeks in Beijing, Taiwan, Amsterdam and Melbourne.
GTC Japan, now in its eighth year, focuses on the key technology trends of AI, robotics and autonomous vehicles. Visitors will attend 70 sessions in eight tracks focused on areas such as autonomous vehicles; deep learning; supercomputing and HPC; professional visualization and VR; and graphics virtualization.
NVIDIA CEO Jen-Hsun Huang kicked off the event by saying that the world was in the midst of an AI revolution – largely driven by GPU computing.
To demonstrate the power of AI, Huang showed a deep learning demonstration in which a real-time artist filter was applied to live video. Using this technology, developed by Artisto, Huang said it’s possible to give video the appearance of famous artists like Dali or van Gogh.
Calling Japan “the epicenter of robotics,” Huang announced a new partnership with FANUC, the world leader in autonomous factories and robotics. He said NVIDIA will help FANUC implement AI in its robots and factory systems, which today are primarily used for building cars in the U.S. and Japan. Thanks to deep learning, the robots will be able to learn on their own, instead of being painstakingly programmed for each function they need to perform.
“The factory of the future will be powered by AI,” Huang said. “Robots will work closely with people, and they will be able to learn by themselves.”
Robots lined the main corridor of the conference, with ingenious creations from than a dozen companies and universities such as Toyota, AeroSense (jointly owned by Sony and ZMP) and the Chiba Institute of Technology. All powered by NVIDIA Jetson, the robots show how deep learning and computer vision can be harnessed to create autonomous robots and drones that help perform vital tasks, from lifting heavy packages to industrial inspection.
As with previous GTCs, the NVIDIA Deep Learning Institute attracted a sell-out audience. More than 300 developers and programmers participated in technical sessions to get started in deep learning, learn about object detection, image segmentation and neural network deployment.
Three more regional GTCs will take place in Washington, D.C., Seoul and Mumbai in the next few weeks.