Apple's research paper on AI useful for computer vision technology

Apple Inc. has been extremely secretive about all its research initiatives, but last month they hinted on changing this for better. So, they have now decided to break their vow of secrecy and Apple researchers have published their first AI paper, which focuses on computer vision technology.

This first public research paper has been pursued by vision expert Ashish Shrivastava and a team of engineers including Tomas Pfister, Oncel Tuzel, Wenda Wang, Russ Webb and Apple Director of Artificial Intelligence Research Josh Susskind. This is for the first time Apple has encouraged and credited its team with a published research paper.

It is path breaking because the company has always prevented its staff from openly publishing their research for the larger community. The paper, titled "Learning from Simulated and Unsupervised Images through Adversarial Training," was submitted for review in mid-November before seeing publication through the Cornell University Library on Dec. 22.

The research paper describes techniques of training computer vision algorithms to recognize objects using synthetic, or computer generated, images. Apple notes that relying completely on simulated images yields unsatisfactory results, as computer generated content is sometimes not realistic enough to provide an accurate learning set.

Apple proposes to use two networks to make this process of vision leraning simpler. It proposes an interaction between the network that trains itself to improve the realism of simulated images (in this case, using photo examples) until they're good enough to disarm a rival "discriminator" network.

Hence, the paper tries to solve the problem of teaching AI to recognize objects using simulated images. Simulated images areeasier to use than photos but, poor for adapting to real-world situations. In order to create a conducive environment for learning, Apple proposes a system of refining a simulator's output through "Simulated+Unsupervised learning."