While many might know Nvidia for its gaming GPUs, a big part of Nvidia’s product line is designed for professional work like deep learning and robotics. Nvidia has now opened up a new lab in Seattle to increase research in the field of robotics.
The lab has 50 researchers with 20 coming from Nvidia and 30 from the University of Washington. The director of the lab, Dieter Fox, is also a member of the University of Washington science faculty.
Seattle was chosen for the rich talent provided by the University of Washington. In addition to that, Microsoft and Amazon’s HQ are in close proximity.
Nvidia’s CEO, Jensen Huang, said, “Because of the University of Washington, because of Microsoft, because of Amazon, this has become one of the great hubs of computer science.”
The aim of the lab is to push robotics research even further. The idea is to develop robots that can assist humans in open-ended environments. Talking to VentureBeat, Fox said,
“The kind of robots that are going to come up that I think are going to be like the next generation of robot, the robot that can perform manipulation task alongside people, which means far more unstructured environment robots, for example, that help people with physical disabilities, robots that could ultimately also be deployed in our homes help us with everyday chores, even with cooking tasks or things like that, but also manufacturing,”
To demonstrate the type of applications Nvidia will be exploring, Nvidia had a couple of demos including the “kitchen manipulator”.
The robot uses deep learning to analyze the place and then track various objects in the scene. The breakthrough here is that it does not need to know the place beforehand. It can manipulate the kitchen environment by opening drawers and perform even more complex task like picking up the pan.
The complexity of the kitchen environment possesses a lot of learning for the robot which can be applied in more advanced situations as noted by Fox.
The past few years have seen Nvidia amp up its research departments. Nvidia’s deep learning team has been producing great results let’s see what their robotics team does.
