Apple AI chief showcased what the company’s self-driving team have achieved so far, which includes several crucial breakthrough technologies.
In sharp contrast to its earlier ultra-secretive approach to its self-driving car project, Apple now seems eager to discuss the progress it has made so far. Unfortunately, from what the AI research director, Ruslan Salakhutdinov had to say, it still is a work-in-progress thing though there are some crucial achievements made so far.
For instance, Apple claims to have developed a system that enables the onboard 3D cameras or lidars to pick up details even in adverse conditions. For instance, the camera will still be able to pick up obstacles even if say raindrops are covering the lens’ surface. Similarly, the system will also be able to estimate the position of a pedestrian even if he or she might be obstructed by a parked vehicle.
Further, to have the right map happens to be one of the most critical aspects of any self-driving car technology and Apple aims to achieve that using both localization and mapping at the same time, something technically referred to as SLAM. This coupled with inputs from the car’s onboard sensors will help it create a detailed 3D map that can be used to suggest directions to the car.
Plus there will be extensive algorithms in place to enable the system to make the right decisions even in crunch situations where there might be the need to eliminate a lot of contesting possibilities. It can be like tackling an errant pedestrian or traffic as a whole in snowy conditions where visibility is severely impaired while having to make through unknown terrain.
Salakhutdinov was speaking to some 200 AI experts when he made the above revelations. Also, while his discourse did provide the world of what it has been up to so far as Apple’s self-driving car project is concerned, it highlights another concern the company is facing as it goes ahead with developing its first autonomous driving car; that of coming across and retaining enough AI talent needed to push ahead with its dreams.
While Apple has so far required its AI experts to maintain a heavy cloak of secrecy around the progress it makes, the same is again contrary to what the experts yearn for, that of showcasing the same to research institutions to help them attain recognition, and perhaps better placements ahead. That Apple is having to yield to their demands shows it isn’t in the position to call the shots, something that it would prefer to have.