[NetEase smart news December 11 news] Apple's progress in the field of machine learning has been emphasized in the past, but in a speech this week, the company's AI executives discussed this issue in more depth. As mentioned in the Wired Magazine, Ruslan Salakhutdinov showed us on stage how Apple used machine learning. Salakhutdinov's speech was presented at a machine learning conference called "NIPS" (Neural Information Processing Systems), which has been ongoing for 31 years and has grown exponentially in recent years, attracting 8,000 people this year. Most of Salakhutdinov's speeches revolved around the impact of machine learning on autonomous vehicle systems. For example, he discussed projects around detecting cars and pedestrians on busy streets, driving on unknown streets, and drawing detailed 3D maps of cities. Salakhutdinov demonstrated Apple's system for identifying cars, pedestrians, and driving areas on the road. He showed how this system works, even in exceptional circumstances, such as when it rains outside, or when pedestrians and other hazards are out of sight. This system can even infer where the pedestrian may be. He quipped: "If you ask me five years ago, I would very much doubt whether you can do it." Another project that Salakhutdinov shows is the foundation of a research paper published by Apple last month. The core of the paper is the use of laser radar to detect pedestrians and cyclists. Salakhutdinov also showed how Apple collected data from cars on the road and used the data to create a wide range of detailed 3D maps to provide traffic lights and various road markings. Such maps play an important role in self-driving cars. Another item discussed by Salakhutdinov is a technique called Simultaneous Localization and Mapping (SLAM), which allows software to move in a certain direction for instant location and map construction. SLAM is used in robots and self-driving cars and has applications in map construction and augmented reality. The fourth item is to generate a 3-D map using the data collected by the sensor-laden car, which contains traffic lights and road sign information. Apple's development of the autopilot system has been surprisingly open. Tim Cook called the autopilot system "very exciting" and said that the automation that Apple is developing is "the mother of artificial intelligence projects." Apple also promoted the reform of auto-driving car testing policy. Recently, a report said that Apple officially tested driverless cars in California. It remains to be seen what kind of success Apple will eventually develop on the autopilot system, but it is certainly interesting to hear its publicity. (Selected from: 9to5mac Compiler: NetEase See Compiler Robot Review: nariiy)