Published: Fri, May 04, 2018
Science | By Joan Schultz

Facebook Training Image Recognition AI with Billions of Instagram Photos

Facebook Training Image Recognition AI with Billions of Instagram Photos

Today at F8, its annual developer conference, Facebook announced it has pioneered a new form of artificial intelligence research demonstrating that training image recognition networks on huge collections of public Instagram photographs produced better image recognition results than traditional, manual annotation of images.

And by training Facebook 's computer vision system with a 1 billion-image version of this dataset, out of the total 3.5 billion Instagram data set, they were able to achieve a record high score of 85.4% accuracy. The company also used Instagram hashtags to train its image-recognizing AI to identify objects more accurately.

Facebook realized it already had a massive public repository of labeled images on its hands: Instagram, which hosts countless billions of photographs, many of which have hashtags, and often multiple tags, which increases the amount of useful information. This is more so because there is no real logic when people use them on the service. While other image recognition benchmarks may rely on millions of photos that human beings have pored through and annotated personally, Facebook had to find methods to clean up what users had submitted that they could do at scale.

The post Facebook utilizes Instagram photos and hashtags to create a smarter A.I. appeared first on Digital Trends.

In the race to continue building more sophisticated AI deep learning models, Facebook has a secret weapon: billions of images on Instagram . It can tell dog breeds, plants, food and plenty of other things that it's grabbed from WordNet.

The accuracy of this data was not an important factor. What is impressive is how the pre-training processes were used to clear out the noise and make the billions of images more useful in order to be used as training data.

According to Wired's report, Manohar Paluri, who leads Facebook's applied computer vision group, said it could be used to identify objectionable or illegal content or describe images to people who are visually impaired.

Like this: