Microsoft's artificial intelligence computing researchers have breakthrough in speed and efficiency with Project Adam. Under the project Microsoft's voice assistant Cortana will be given the ability to identify various objects.
Project Adam's team claims that it is 50 times faster and twice as accurate as Google's 16,000 computer neural network-powered image recognition system. Microsoft Research's executive vice president of technology and research, Harry Shum, showed off the technology at the company's Faculty Summit. Cortana was successfully able to recognize several dog breeds from just a photograph. The team added that in future the system could recognize foods by looking at its image and get its nutritional information, or images of insects could be classified as poisonous or not. The system could also help build new tools for visually challenged users.
"Imagine if you could help blind people see by pointing a cellphone at a scene and having it describe the scene to them Imagine if you could help blind people see by pointing a cellphone at a scene and having it describe the scene to them," said team leader Trishul Chilimbi, describing how such a system could eventually be of use to end-users, in a statement. "We could do things like take a photograph of food we’re eating and have it provide us with nutritional information. We can use that to make smarter choices."
Chilimbi adds that he believes the system is better than other rivals because of the unique way in which it handles the data being processed across its servers. The research team created a "big data" image set, with over 14 million unique images in 22,000 categories. Project Adam established relationships between data points, which is similar to neurons in a human brain.
The "neural network" develops when the machine begins to automatically recognize patterns and build its own framework with which to understand new images. Project Adam is has been called a "deep neural network", which uses multiple layers or hierarchies of such capabilities.
"One of the innovations we came up with was saying that not only can we make it asynchronous, but we went whole hog and decided not to pretend it's synchronous in any way," said Chilimbi. The asynchrony also helps us escape from ruts where the task accuracy does not improve much ... Much like how humans learning a new task often find themselves plateauing after a period of rapid improvement."
Microsoft also plans to integrate Cortana with a academic data into Bing, as part of the "Microsoft Academic Search" project which will give the voice assistant quick access to data. The company is planning to develop a community portal for academic workers, where the researchers can control how much of their personal data is visible.