Apple has provided an insight into its AI research at an invite-only meeting at Tuesday’s Neural Information Processing Systems conference.
As reported by Quartz, the company’s new head of machine learning, Russ Salakhutdinov, along with other Apple employees detailed how Apple is using AI to solve several problems.
Among the areas of research being conducted is machine learning, which involves developing AI that can recognise images, predict user behavior and events in the real world, and developing virtual assistant language.
Related: Apple Car
Machine learning is also being investigated by other companies conducting AI research, most notably Google with its DeepMind division.
Apple reportedly claimed during its presentation that its GPU-based implementation of image recognition could process twice as many photos per second as Google’s system.
Slides provided to Quartz seem to indicate Apple is following numerous other research paths which other companies are also investigating.
That includes ‘volumentric detection of LiDAR’, i.e. measuring and identifying objects with lasers – a technology used for autonomous systems in vehicles.
This doesn’t necessarily mean Apple is working on the tech for a self-driving car, as LiDAR is also used in map-making, and, according to Quartz, two presentation attendees stressed Apple made no mention of cars.
Related: CES 2017
But this week Apple publicly revealed its interest in driverless cars in a letter written to America’s National Highway Traffic Safety Administration.
The letter describes Apple’s ongoing interest in self-driving vehicles, explaining: “Apple uses machine learning to make its products and services smarter, more intuitive, and more personal.
“The company is investing heavily in the study of machine learning and automation, and is excited about the potential of automated systems in many areas, including transportation.”
Elsewhere, it looks like the company is working on building neural networks 4.5 times smaller than the originals but with the same accuracy and twice the speed.
The technique involves using a larger neural network to teach another, smaller, network about the possible decisions it could make in specific situations.
This results in the smaller network essentially being able to predict the decisions of the larger network, allowing for the technology to be included in much smaller devices.
WATCH: Tesla Model S Review – The Ultimate Rental Car?
Let us know what you think of Apple’s AI ambitions in the comments.