Earlier this December, Apple created a noise when they express the artificial intelligence (AI) research community that the secretive company will start making AI papers of its own. A month later, Apple is already starting to keep their promise.

Apple revealed its very first AI paper last December 22. The AI paper was submitted last November 15 for publication. The document represents a technique for how to enhance the training of an algorithm's ability to notice images with the use of computer-generated images rather than real world images.

Using synthetic images in machine learning research is similar to creating it for video games. The technique is to train neural networks, which are more efficient than using real world images. It is because synthetic image data is already marked and expounded, while real world image data depends upon something to exhaustively label everything the computer is programmed.

"The synthetic image data is sometimes not realistic enough, bringing the network to gain details only present in synthetic images and fail to generalize well on real images," the paper said.

Enable to develop the training with synthetic image data, the paper advices what the Apple reseachers named "Simulated+Unsupervised" learning, where the simulation of a real image is boosted. They used a modified version of a new learning machine technique called "General Adversarial Networks," which compared two neural networks between each other.

Researcher Ashish Shrivastava is the lead author of the Apple AI paper, who has a PhD in computer vision from University of Maryland, College Park. Other Apple employees were co-authors on the paper including Tomas Pfister, Oncel Tuzel, Wenda Wang, Russ Webb and Josh Susskind, who also co-founded an AI startup that determines a person's emotions by looking at facial expressions they called "Emotient," an Apple-made application which was made this year.

The first AI Paper sets a major achievement for the Apple to move ahead more with the community. AI software is becoming fundamental for everything from the camera capabilities in Apple's latest iPhones to the internet services functioning in the data center.