By Mike Schroepfer, Chief Technology Officer 

Today we’re announcing new milestones achieved in Facebook’s long-term artificial intelligence research. The advancements come from Facebook’s AI Research (FAIR) team. They include the ability to train computers to identify objects in photos, understand natural language, predict, and plan.

Object detection
Next month FAIR will be presenting a new paper at NIPS, a leading artificial intelligence conference. In the paper the team details a state-of-the-art system that segments, or distinguishes between, objects in a photo. This new system segments images 30 percent faster, using 10x less training data, than previous industry benchmarks.

Screen Shot 2015-11-02 at 10.15.18 PM

Natural language understanding meets image recognition
Earlier this year, we showed some of our work on natural language understanding — specifically, a system called Memory Networks (MemNets) that can read and then answer questions about short texts. In this demo of a new system we call VQA, or visual Q&A, MemNets are combined with image recognition technology, making it possible for people to ask the machine what’s in a photo.

Predictive learning
Unsupervised, or predictive learning, is the ability to understand what will happen in the future by learning from observation. To try to give computers this ability, the FAIR team has developed a system that can “watch” a series of visual tests — in this case, sets of precariously stacked blocks that may or may not fall — and predict the outcome. After just a few months’ work the system can now predict correctly 90 percent of the time, which is better than most humans.

To teach systems how to plan, FAIR has created an AI bot to play the board game Go. After a few months of playing, it’s already on par with the other AI-powered systems that have been published and it’s as good as a very strong human player. We’ve achieved this by combining the traditional search-based approach — modeling out each possible move as the game progresses — with a pattern-matching system built by our computer vision team.

Our AI research efforts — along with our work to develop radical new approaches to connectivity and our work to develop immersive new VR technologies — are a long-term endeavor. But if we can get them right we will be able to build systems that are smarter and more useful, enable developers to create immersive new experiences, and make it possible to connect everyone in the world.

To learn more about how we’re approaching AI research and the impact it’s already having, check out this video.