Microsoft Debuts Project Adam Object Recognition AI

The company unveiled a new deep-learning system that aims to mimic the human brain's ability to quickly make sense of the world around it.

Microsoft machine learning

Microsoft earlier this year teased that the company's innovations in artificial intelligence (AI) are inching closer toward commercialization. Now, the company is offering a glimpse of that future in the form of Project Adam.

Making its debut at the Microsoft Research Faculty Summit, Project Adam is a "deep-learning" system patterned after the human brain's network of neurotransmitters. According to Microsoft, the system can recognize objects 50 times faster with twice the accuracy of other systems. The project's aim is "to enable software to visually recognize any object," said the company.

Researchers collected 14 million images—organized into 22,000 categories via user-generated tags—from Flickr and other online sources. The data was then "used to train a neural network made up of more than two billion connections" using 30 times fewer machines than competing systems.

During a live demonstration, Johnson Apacible, a Project Adam researcher, showed off how the technology can help users identify things, including the breeds that make up man's best friend.

Apacible asked Cortana, Microsoft's voice-enabled digital assistant for Windows Phone 8.1, to determine what type of dog was brought on-stage. A moment after snapping a picture, the Project Adam-enhanced app correctly identified the pooch as a Rhodesian Ridgeback.

Cortana also correctly identified a Dalmatian, before bumping into the limits of AI. The dog breed app identified another dog as a terrier, while the audience pegged it as a Labradoodle. They were right and wrong in that both breeds are found in the Cobberdog, the correct answer.

While it's a novel use of the technology, at least for dog lovers, Microsoft envisions bigger things for Project Adam.

Harry Shum, executive vice president of Microsoft Technology and Research, stated that the "computing paradigm is shifting from personal computing of the past to the future, as the user is now in the center." Rather than focus on "computing power or storage or bandwidth," he asserted the new focus is "about people's time and attention."

The tech behind Project Adam could one day make it possible for users to snap pictures to diagnose skin conditions or determine the nutritional makeup of a meal. "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," boasted the company.

Microsoft isn't the only tech giant that is pursuing advanced AI.

Last year, Google launched the Quantum Artificial Intelligence Lab to study how quantum computing could advance the field of AI. "Machine learning is all about building better models of the world to make more accurate predictions. ... And if we want to build a more useful search engine, we need to better understand spoken questions and what's on the web so you get the best answer," stated Hartmut Neven, director of engineering for the Google Research team, at the time.

In January, Google announced that it had snapped up London-based DeepMind for $400 million in cash. The search giant reportedly outbid Facebook in acquiring the AI software maker.

Pedro Hernandez

Pedro Hernandez

Pedro Hernandez is a contributor to eWEEK and the IT Business Edge Network, the network for technology professionals. Previously, he served as a managing editor for the network of...