This article was published on October 16, 2017

Google’s AI can create better machine-learning code than the researchers who made it


Google’s AI can create better machine-learning code than the researchers who made it Image by: Arthur Caranta

Google’s AutoML system recently produced a series of machine-learning codes with higher rates of efficiency than those made by the researchers themselves. In this latest blow to human superiority the robot student has become the self-replicating master.

AutoML was developed as a solution to the lack of top-notch talent in AI programming. There aren’t enough cutting edge developers to keep up with demand, so the team came up with a machine learning software that can create self-learning code.  The system runs thousands of simulations to determine which areas of the code can be improved, makes the changes, and continues the process ad infinitum, or until its goal is reached.

This is a fabulous representation of the infinite monkey theorum, but instead of a monkey with a keyboard creating Shakespeare, Google made machines capable of replicating their own programming. And those machines can do in hours what takes the best human programmers weeks or months.

Even scarier, AutoML is better at coding machine-learning systems than the researchers who made it. In an image recognition task it reached record a high 82 percent accuracy. Even in some of the most complex AI tasks, its self-created code is superior to humans; it can mark multiple points within an image with 42 percent accuracy compared to human-made software’s 39.

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!

This isn’t the beginning of Skynet or anything spooky like that, we’re not on the verge of the singularity (self-aware machines), but we are leaps closer to revealing AI’s potential to accelerate the technology timeline.

Google only announced AutoML five months ago. It’s remarkable the company created an AI capable of creating better machine-learning systems than researchers have, in such a short time – imagine where the same team will be in a year. More, the team is likely to share their research with others both inside and outside of the company, as Google’s AI researchers often do.

I can’t be the only person who wants to see what Deep Mind can do with the ability to create a better Deep Mind — a Deeper Mind, if you will?

The AI revolution is currently underway and its future is brighter than ever. AutoML is the genesis for the next generation of machine-learning. Tomorrow’s machines won’t just learn, they’ll self-update and be capable of creating custom programs to solve unforeseen problems.

Our hope lies in a future where AI takes care of time-consuming tasks like programming, thus freeing humans to do the things machines can’t do – like enjoy tacos and beer.

Get the TNW newsletter

Get the most important tech news in your inbox each week.

Also tagged with


Published
Back to top