Earlier this month word spread about a change in Apple's policy regarding artificial intelligence research. In line with its reputation for secrecy, Apple historically has not allowed employees to publish their research, which many have speculated could make the company a less attractive workplace for AI researchers. But Quartz reported that Russ Salakhutdinov, a director of AI research at Apple, claimed research would soon begin to be published, and a greater effort made to work with the broader research community. The first fruits of that claim were uncovered this week, as Mitchel Broussard of MacRumors reported on the first research paper being published:
Titled "Learning from Simulated and Unsupervised Images through Adversarial Training," the paper describes a program that can intelligently decipher and understand digital images in a setting similar to the "Siri Intelligence" and facial recognition features introduced in Photos in iOS 10, but more advanced.
The biggest news here is not in the research paper itself, but in what it represents for Apple going forward: newfound openness in a subject that will likely become increasingly more important in the years to come.