Instagram, the Facebook owned company contains photo of an exquisitely framed sunset which is helping Facebook train its artificial intelligence algorithms to better understand objects in images, the company announced at its annual F8 developer conference.

Facebook claims the approach is the best way to amass and train software with billions of images without requiring for human input to analyze the data or annotate it.

According to Mike Schroepfer, Facebook’s chief technology officer, “We rely almost entirely on hand-curated, human-labeled data sets. If a person hasn’t spend the time to label something specific in an image, even the most advanced computer vision systems won’t be able to identity it.”

Thus Facebook collected relevant data and use it to train its computer vision and object recognition models. And the end result is a training system that created algorithms Facebook believe beats top-of-the-line industry benchmarks.

Facebook posses extremely valuable text and image data it can use to inform its AI models, so long as that text and those images are posted as public by its users.

Albeit, most of the users may not be aware that the public data they shared are being mined to build AI systems, and not just for the already known reason of serving ads, which concerns may be farfetched, given that such data can not constitute any privacy issue.

Facebook only extract object-based data, and it’s not necessarily trying to draw inferences about user behavior from the contents of the photos.

How Facebook employ Instagram images to train artificial intelligence (AI) algorithms



Instagram, the Facebook owned company contains photo of an exquisitely framed sunset which is helping Facebook train its artificial intelligence algorithms to better understand objects in images, the company announced at its annual F8 developer conference.

Facebook claims the approach is the best way to amass and train software with billions of images without requiring for human input to analyze the data or annotate it.

According to Mike Schroepfer, Facebook’s chief technology officer, “We rely almost entirely on hand-curated, human-labeled data sets. If a person hasn’t spend the time to label something specific in an image, even the most advanced computer vision systems won’t be able to identity it.”

Thus Facebook collected relevant data and use it to train its computer vision and object recognition models. And the end result is a training system that created algorithms Facebook believe beats top-of-the-line industry benchmarks.

Facebook posses extremely valuable text and image data it can use to inform its AI models, so long as that text and those images are posted as public by its users.

Albeit, most of the users may not be aware that the public data they shared are being mined to build AI systems, and not just for the already known reason of serving ads, which concerns may be farfetched, given that such data can not constitute any privacy issue.

Facebook only extract object-based data, and it’s not necessarily trying to draw inferences about user behavior from the contents of the photos.

No comments