Artificial Intelligence
Facebook's AI research team is working to build smart systems that can enhance people's lives. The team recently shared their AI research and the impact this work is already having. |
With an ever-exploding number of posts to Facebook newsfeeds by the site's over 1 billion users, there is no way human-based systems can hope to catch up with or even keep track of the social networks user collateral. Heading up this challenge, Facebook AI Research (FAIR) has been conducting ambitious work in areas like image recognition and natural language understanding. They've already published a series of groundbreaking papers in these areas, and recently announced a few more milestones.
The research includes a system that can analyze photos, determine what they depict, and offer a verbal description—a feature that is particularly valuable for visually impaired people.
The researchers have been working to train systems to recognize patterns in the pixels so they can be as good as or better than humans at distinguishing objects in a photo from one another and then identifying each object. The process is known as "segmentation."
Teaching machines to see and understand
Facebook's AI team is working to build smart systems that can enhance people's lives. Watch this video to learn about how we're approaching AI research and the impact this work is already having.
Posted by Facebook Engineering on Tuesday, November 3, 2015
| Related articles |
MemNets add a type of short-term memory to the convolutional neural networks (CNN) that power FAIR's deep-learning systems, allowing those systems to understand language more like a human would. "Now we've scaled this system from being able to read and answer questions on tens of lines of text to being able to perform the same task on data sets exceeding 100K questions, an order of magnitude larger than previous benchmarks," report the researchers.
They have also made technological leaps in predictive analysis and unsupervised learning, The FAIR team recently started to explore these models, and 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.
The team claims to also be making progress in the ever-challenging AI problem of the game Go. "We’ve been working on our Go player for only a few months, but it's already on par with the other AI-powered systems that have been published, and it's already as good as a very strong human player," they claim.
"Because now that it works, there are so many doors that are open all of a sudden.." |
LeCun says artificial intelligence now has real potential for sustained gains. "Now computer vision and speech recognition just work," he said. "They’re not perfect, but they work. And that enables a lot of applications, which is why you see all this excitement around deep learning and AI. Because now that it works, there are so many doors that are open all of a sudden."
According to resarcher Mike Schroepfer, "Our AI research — along with our work to explore radical new approaches to connectivity and to enable immersive new shared experiences with Oculus VR — is a long-term endeavor. It will take a lot of years of hard work to see all this through, but if we can get these new technologies right, we will be that much closer to connecting the world."
SOURCE Facebook
| By 33rd Square | Embed |
Visual Question and Answering Demo
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 our image recognition technology, making it possible for people to ask the machine what's in a photo.
Posted by Facebook Engineering on Tuesday, November 3, 2015
>
0 comments:
Post a Comment