Forget Google, it's Microsoft that's building Sky.NET!
If you watched the Google I/O keynote this year, then you will have witnessed the gentleman at the back shouting about how the web giant is building robots which will kill us all. With Elon Musk also adding his concerns to such a scenario, any potential for a Terminator-style "Skynet" artificial intelligence which could be dangerous to our very existence is under close-observation.
Today we've got a step closer to such a self-thinking AI, but it's not from Google, it's in-fact from Microsoft. So why should we be keeping a close eye on 'Project Adam' from the Redmond giant? Well, drum roll please... the company's technology can now detect puppies. But wait, there's more, the system can specifically differentiate breeds of puppies. Where is Sarah Connor?!
Of course, dogs were only used as a demonstration for Project Adam when Microsoft Research unveiled it on Monday. It shows a powerful system reliant on "deep-learning" which Microsoft claims is "50 times faster—and more than twice as accurate" as Google's system. Microsoft posed the question "Can you tell the difference between the two breeds of corgis?" which, if you're like me, I couldn't. It shows how the machines are already smarter than us.
Many wrongly see Bing as a search engine living in the shadow of Google. Behind both of the search engines are powerful "knowledge graphs" linking the vast amount of information together available around the web to be used for applications such as AI. Bing's arm for this is called 'Satori' and it's nothing short of incredible.
Satori is behind Microsoft's recently-released AI for Windows Phone, Cortana. Named over the artificial intelligence in the 'Halo' franchise, Cortana offers a more personal experience from competitors due to its machine-learning through the collection of user information and individual preferences. To ensure privacy, everything held about the user is stored in an on-device 'Notebook' which individual data can be deleted from.
For some perspective around the power of Project Adam, Google's own project - albeit a couple of years ago - used a network of 16,000 computers to "teach" itself how to identify images of cats. As mentioned earlier, this was also slower and less accurate than what Microsoft achieved using 30x less machines.
“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.
Google has been very vocal in its ambitions to be a leader in deep-learning. The company has even employed the field's pioneer, Geoff Hinton from the University of Toronto, to head research for Google's products. Despite this, it looks like Microsoft is currently the one to watch for Skynet.
In fact, the company says it defied the odds: “There was a lot of skepticism when we started out from machine-learning experts around using the distributed system to do machine learning,” Chilimbi says. “The fundamental machine-learning training algorithms are synchronous. They’ve typically been run on a single machine. They said, “Yes, you can do this distributed, but the synchronization cost will make it so slow that it’s never going to be high-performance or scalable."
“One of the innovations we came up 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. We figured out a way to make the asynchrony not just learn but learn better, because it adds a level of robustness. Learning is not so much about optimizing on the training set of data. It’s about generalizing well on unseen data.”
Even if Microsoft leads us to a world dominated by killer machines, it will be awesome, we'll know every breed of animal imaginable.
Do you plan to utilise machine-learning and/or AI for your applications? Let us know in the comments.