After Apple published research on 3D object detection for autonomous cars, now a company executive has revealed some more details about the system that is being developed for vehicles.
Ruslan Salakhutdinov, Apple's director of artificial intelligence, spoke a little about what is being developed by the company during the machine learning conference, NIPS (Neural Information Processing Systems), as WIRED.
A superb example cited by him is the possibility that the company's camera can already identify objects on the streets even if the lenses are obstructed by rain and also identify pedestrians on the sidewalks even when they are partially hidden behind vehicles, things that, as Salakhutdinov commented, no would have been possible five years ago.
In addition to these new features, the system is also able to make decisions dynamically, such as avoiding pedestrians on the street and also using ?SLAM? (simultaneous localization and mapping, or simultaneous location and mapping) to improve the sense of direction.
The director also said that Apple would create 3D maps of cities, which would include the identification of traffic lights, road signs and more maybe that data is being collected by the Apple vehicle that is walking the streets of the United States (and other countries).
For now, of course, Ma has not yet released what it plans to do with the system once it is ready. But even though she said she would be working on a * system * for autonomous cars, many still have the hope that this is just a gateway to the development of hardware, that is, a car completely created Apple.
As always, we can only wait.