Apple is now publishing a ‘journal’ with its cutting-edge machine learning research

Ruslan Salakhutdinov Apple AI
Apple Director of AI Ruslan Salakhutdinov Screenshot/YouTube

Apple launched a new blog on Wednesday where it plans to publish academic-style papers looking at new applications and techniques in machine learning. 

Machine learning is a buzzword in the technology industry, but it’s also one of the most closely-watched fields of research at the moment. Machine learning is the study of how to program software and algorithms to be able to learn and improve their forecasts and predictions from large sets of data.

“Here, you can read posts written by Apple engineers about their work using machine learning technologies to help build innovative products for millions of people around the world,” the introductory blog post states

The site was announced by Ruslan Salakhutdinov, a respected academic Apple hired last year and Apple’s current director of AI research. 

Critics have said that Apple has fallen behind in machine learning and a related field, artificial intelligence. For example, its Silicon Valley rival, Google, has a subsidiary, DeepMind, which is one of the leaders in a machine learning technology called deep learning or neural networks.

Apple has argued that it uses less advanced machine learning techniques all over its products. Apple CEO Tim Cook has cited a way that Apple uses machine learning to increase how long a battery lasts as his primary example.


Machine learning is also a key technology that will enable self-driving cars, smartglasses, and virtual personal assistants. 

One reason why Apple may have seemed like it was falling behind in machine learning is that it is a secretive company and it previously did not allow its researchers to publish in academic journals, like machine learning specialists at Google and Facebook were allowed to do.

Increasingly, interesting applications and techniques for machine learning have been published on ArXiv, a website that collects academic-style papers before they have been peer-reviewed. A few papers from Apple researchers have been spotted on ArXiv, but going forward some may end up on this new Apple blog.

One reason why Apple may have started this blog is because papers on ArXiv are supposed to share their training dataset, which Apple doesn’t want to do. 

A graphic from the Apple paper Apple

The first substantive post examines a way to generate simulated images to train a machine learning algorithm using a technique called generative adversarial networks. It appears as if the papers published on the blog will not name the Apple employees that authored them.