Microsoft's Project Adam AI Capable of Identifying Different Dogs

Microsoft Corp. demonstrated a new system for artificial intelligence this week called Project Adam, which is capable of distinguishing between different breeds of dogs.

The Redmond, Washington-based company said in a blog post that the project is aimed toward giving software the ability to visually recognize any object, according to PC Magazine.

"It's a tall order, given the immense neural network in human brains that make those kinds of associations possible through trillions of connections," Microsoft wrote. The company added that the system, which was shown off at this week's Microsoft Research Faculty Summit, "exhibits world-class results in performance, scaling, and accuracy for visual-recognition tasks."

During the demonstration, Adam used computer vision and insights from huge amounts of data to automatically tell the difference between breeds of dogs in real time, GeekWire reported. Microsoft said the technology is able to improve over time thanks to its use of an asynchronous technique.

Microsoft is among the many tech companies working on AI systems, claiming in its post that the project surpasses machine learning endeavors from other companies such as Google, who worked in the past on using the technology to identify different cats.

The goal of Project Adam is to create a computer with the same complex processes of the human brain, giving it the ability to understand requests the same way a person can.

While demonstrating the system, a Microsoft researcher pointed a Windows phone at a dog. The researcher then asked Cortana, powered by Adam, to identify the breed, to which Cortana said it was a Rhodesian Ridgeback. The system was also able to identify a Dalmatian and a Cobberdog, and when the phone was pointed at Harry Shum, executive vice president of Technology and Research, the system said, "this is not a dog."

"Using 30 times fewer machines than other systems, that data was used to train a neural network made up of more than two billion connections," Microsoft said. "This scalable infrastructure is twice more accurate in its object recognition and 50 times faster than other systems."

While Project Adam is still in its early stages of development, its creators are looking to give the system other abilities, such as letting users point their phones at food to find the nutritional value in meals, PC Magazine reported.

"Or, snap a photo of an unusual skin condition and get an accurate diagnosis. Or if you're out in the woods and need to know which plants are poisonous and which are edible, this is the technology that could help you do that," Microsoft said.

Tags
Microsoft
Real Time Analytics