Apple’s New Machine Learning Blog Hints at AI Ambitions

Apple's New Machine Learning Blog Hints at AI Ambitions
Text Size
- +

Toggle Dark Mode

Apple on Wednesday debuted a new platform where interested users can learn about the company’s machine learning endeavors from the engineers themselves.

That platform is a simple blog-style website titled “Apple Machine Learning Journal.” Currently, the site is host to a welcome message and a single post — but will presumably continue to grow in the coming months and years. In the welcome post, Apple encourages machine learning researchers, students, engineers and developers to email questions and feedback to Apple’s team at

The first blog post, simply titled Vol. 1, Issue 1, focuses on improving the image recognition capabilities of neural networks. On one hand, it’s preferable to use a larger, diverse and accurately annotated library of real images to “train” neutral nets — but this is costly. A better and more cost-effective alternative is to use synthetic images. Of course, the downside is that the synthetic images may be of poorer quality and not realistic enough to properly train a neural net. To counteract that, Apple’s engineers have developed a way to refine synthetic images to be much more realistic and useful for training machine learning platforms. Eagle-eyed followers of Apple’s endeavors will recognize this advancement as the subject of the company’s first published AI research paper, which was released in December 2016.

Apple is one of the foremost researchers of artificial intelligence and machine learning but has historically kept its developments secret due to concerns over losing its commercial intellectual property. That all changed last year when Apple shifted its policy to allow for its team to publicly release their research. That policy change and this new machine learning blog are not only signs of a newfound transparency but are also indicative of Apple’s efforts to bump up its research capacity. According to Bloomberg, publication and presence are key ways to boost recruiting of qualified graduates.

As for other signs that Apple is doubling down on AI, earlier this year, the Cupertino company became a founding member of the Partnership on AI, a nonprofit research initiative whose ranks include some of the world’s biggest tech giants. In addition, Apple is significantly expanding up its AI and machine learning initiatives in Seattle, Washington — a city that’s quickly becoming a hub for such research. Apple also holds a number of patents and acquisitions related to artificial intelligence and machine learning.

And while it all sounds pretty cerebral, Apple’s efforts in the AI sphere could have a profound impact on everyday consumers. When CEO Tim Cook announced a new machine learning and artificial intelligence R&D center in Japan last October, he hinted at the myriad of ways that AI could improve the company’s flagship iPhone, from better battery life management to significantly improved content recommendations.

Social Sharing