NVIDIA for the future!

published on February 15, 2018 on voicesofvr.com, listened on March 3, 2018

In this Voices of VR podcast, Omer Shapira, a senior VR designer and engineer at NVIDIA, talks about training artificial intelligence and robots in VR. Shapira’s main focus is designing human aspects of VR interactions. Last year, NVIDIA came out with something they call Project Holodeck which is a VR environment that mimics the real world through sight, sound, and haptics. It even gives the user “hands” and full dynamic control in the space. Its potential is endless. For instance, at a demo, NVIDIA took an audience inside the design of a Koenigsegg Regera supercar. This allows the audience to view the car at full scale and see all of its components. At SIGGRAPH, NVIDIA showed how VR can be used to train AI and robots in real time rendering and realistic test scenarios. There is also no risk since it’s all virtual! In the demo, there was a robot working with dominoes and in another room, there was a VR headset where a person could train the robot how to play with the dominoes. There is another possibility of having one machine who knows how to play teach another machine through reinforcement learning, similar to how one teaches a child. This kind of reminded me of

Neo in The Matrix where he learns kung fu once the information is inputted in his system/mind. Shapira makes it apparent that the more we interact with AI and robots, the more we should be concerned that they are programmed well and be confident that they will do the right thing. I’ve never thought about this before but it does make sense. I guess I always trusted the ones that created the technology will create something that performs its intended purpose while being safe. Now I see how that is a lot to ask. Shapira continues and tells us what he hopes NVIDIA’s Holodeck will be used for such as training robots to help the disabled efficiently and effectively. I wonder if we will have to train robots specifically for a person’s personality as well, but that’s for a later time.

Related posts:

Getting you to speed with the Basics of PyTorch.