Apple machine learning efforts have been highlighted in the past, but the company’s artificial intelligence director dove deeper during a speech this week. As outlined by Wired, Ruslan Salakhutdinov took the stage to offer a closer look at how Apple is using machine learning…
Salakhutdinov’s speech was at a machine learning conference called NIPS, which has been held for the last thirty-one years and has grown exponentially over recent years, attracting 8,000 people this year.
Much of Salakhutdinov’s talk centered around the implications machine learning has on self-driving car systems. For instance, he discussed projects centered around detecting cars and pedestrians on busy streets, navigating unknown streets, and building detailed 3D maps of cities.
Salakhutdinov demonstrated Apple’s system for identifying cars, pedestrians, and drivable parts of the road. He showed how the system works even when it’s raining outside or when pedestrians and other hazards aren’t in direct line of sight. The system can even infer where a pedestrian might be. “If you asked me five years ago, I would be very skeptical of saying ‘Yes you could do that,’” he quipped.
Another project demonstrated by Salakhutdinov was the basis of a research paper published by Apple last month, centered around using lidars to detect pedestrians and cyclists.
Salakhutdinov also showed how Apple is collecting data from cars on the road and using that data to build extensive and detailed 3D maps, offering information such as traffic lights and various road markings. Such maps are instrumental in operating autonomous vehicles.
Another project Salakhutdinov discussed involved giving software moving through the world a kind of sense of direction, a technique called SLAM, for simultaneous localization and mapping. SLAM is used on robots and autonomous vehicles, and also has applications in map building and augmented reality.
A fourth project used data collected by sensor-laden cars to generate rich 3-D maps with features like traffic lights and road markings.
Apple has been surprisingly open about some of its work in creating autonomous systems. Tim Cook has called autonomous systems “incredibly exciting” and said Apple is working on the “mother of all” AI projects. Apple has also pushed for changes to self-driving car testing policy. Most recently, a report said that Apple was officially testing self-driving cars in California.
What ultimately comes of Apple’s efforts in autonomous systems remains to be seen, but it’s certainly interesting hearing the company be so open about it. Read more at Wired.