Where Google's Neural Networks Recognized Cats, Microsoft Sees Dogs, and Faster

 Artificial Intelligence  

Microsoft has just upped the ante for artificial intelligence.  Project Adam is a new deep-learning system modeled after the human brain that has greater image classification accuracy and is 50 times faster than other systems in the industry.

Microsoft Research has developed a new artificial intelligence system using machine learning, called “Project Adam." The software increases the speed and efficiency of computers and their ability to learn.  Project Adam uses inspiration from the human brain to absorb new data and teach itself new skills — such as distinguishing among different breeds of dogs.

Related articles Dennis Hong On Robot Evolution and The DARPA Robotics Challenge  Boston Dynamic's Big Dog Gets An Arm For A Head  The What, Why, How, Who and When of the DARPA Robotics Challenge

Project Adam aims to demonstrate that large-scale, commodity distributed systems can train huge deep neural networks effectively. For proof, the researchers created the world’s best photograph classifier, using 14 million images from ImageNet, an image database divided into 22,000 categories.

The system was demonstrated at Microsoft’s Faculty Summit in Redmond recently (video below), as Microsoft brought out several different breeds of dogs on stage and showed how the technology could automatically distinguish among them in real time, using computer vision and insights from large sets of data.  The system was integrated into Cortana, Microsoft's digital assistant platform.

Microsoft says Project Adam has achieved breakthroughs in machine learning by using distributed networks and an asynchronous technique that improves the overall efficiency and accuracy of the system over time. This is a critical area of technology as Microsoft and other companies race to build intelligent, predictive systems that leverage mobile technologies and the cloud.

"Project Adam knows dogs. It can identify dogs in images. It can identify kinds of dogs. It can even identify particular breeds, such as whether a corgi is a Pembroke or a Cardigan."

Now, if this all sounds vaguely familiar, that’s because it is—vaguely. A couple of years ago, Google used a network of 16,000 computers to teach itself to identify images of cats. That is a difficult task for computers, and it was an impressive achievement.

"We wanted to build a highly efficient, highly scalable distributed system from commodity PCs that has world-class training speed, scalability, and task accuracy for an important large-scale task."

According to Microsoft Research, Project Adam is 50 times faster—and more than twice as accurate, as outlined in a paper currently under academic review. In addition, it is efficient, using 30 times fewer machines, and is scalable, areas in which the Google effort fell short.

“We wanted to build a highly efficient, highly scalable distributed system from commodity PCs that has world-class training speed, scalability, and task accuracy for an important large-scale task,” says Trishul Chilimbi, one of the Microsoft researchers who spearheaded the Project Adam effort. “We focused on vision because that was the task for which we had the largest publicly available data set.

“We tend to overestimate the impact of disruptive technologies in the short term and underestimate their long-term impact—the Internet being a good case in point. With deep learning, there’s still a lot more to be done on the theoretical side," Chilimbi says.



SOURCE  Microsoft Research

By 33rd Square

Embed