Your #sunset photos might have helped improve Facebook's AI

Christopher Davidson
May 4, 2018

Since Facebook owns Instagram, it is no surprise that the company used the photo-sharing service's big image database to improve the platform's image recognition capabilities.

Facebook's director of AI and machine learning, Srinivas Narayanan, explains artificial intelligence as it relates to image recognition and categorization.

From its early days, hashtags have been one of the most fundamental parts of content sharing on any social media.

Interestingly, Facebook's AI scientists have found a different use for your hashtags.

During the opening F8 2018 keynote, Facebook CEO Mark Zuckerberg showed off the company's latest Instagram updates: Spotify integration, AI-based anti-bullying comment filters, AR camera effects and four-way video chat.

In the race to continue building more sophisticated AI deep learning models, Facebook has a secret weapon: billions of images on Instagram . The increases in image recognition accuracy only were a couple of points in numerous tests, but what's fascinating are the pre-training processes that turned noisy data that was this vast into something effective while being weakly trained. The work beat the previous state-of-the-art model, which utilized supervised, hand-labeled data, by 2%.

Missouri Legislature Calls Special Session to Consider Impeaching Governor Eric Greitens
His criminal trial on felony invasion of privacy charges begins May 14, four days before the special session. In the past, special sessions have only been called by the governor.

Why manually label images when you already have a source for millions of images labeled with hashtags? This is more so because there is no real logic when people use them on the service.

Using this strategy, called "weakly supervised training", Facebook's AI achieved a record 85.4% accuracy rating on an industry-wide test of image recognition, beating out Google's previous record.

However, the messy and unorganised hashtagging habits of Instagram users did cause them some troubles.

On the other hand, having multiple hashtags on an image also came with a surprising advantage.

Facebook's chief technology officer Mike Schroepfer unveiled that the company's computers are now more accurate when trying to detect specific objects in users' visual input.

Other reports by

Discuss This Article