What it is: Apple released a new software framework to enable machine learning.
Machine learning has taken the promise of artificial intelligence and made it practical. In the early days of artificial intelligence (AI), programmers had to explicitly write code to mimic intelligence. That meant anticipating every possible outcome, which made early AI projects like expert systems ultimately unwieldy and impractical. Machine learning simplifies AI by defining rules that the computer follows to learn from real-world data.
Credit card companies have been using machine learning to identify possible fraud in credit cards by noticing unusual patterns. If you live in Dallas, Texas but the credit card company suddenly notices purchases made in Miami, Florida and then another in Portland, Oregon that same day, that’s clear something could be wrong.
Spam filters also use machine learning to identify possibly spam and get better at filtering it out. As you identify junk mail, the spam filter learns what you consider spam and gets better at detecting and blocking it ahead of time.
Now machine learning is poised to break through to common computing tasks. Google is adding machine learning to their spreadsheet and Apple is applying machine learning to vision recognition and natural language processing. Vision recognition will help identify faces and objects in a picture while natural language processing will help Siri learn from spoken commands and allow virtual keyboards to offer suggestions based on what you’ve been searching or typing recently.
By offering machine learning as a software framework, Apple has made it easy for developers to add machine learning capabilities to their own apps. Apple is also rumored to be developing a dedicated neural network processor. Combine this neural processor with their machine learning framework and you can see that in the near future (think this fall with the new iPhone), Apple products will include machine learning capabilities that will surpass rival products that still depend on the cloud (Internet) for machine learning tasks.
Apple is moving machine learning to the device itself. One useful example will be language translation. Siri can translate English to several languages including French, Italian, German and Mandarin. It’s only a matter of time before Siri will be able to translate other languages into English as well. By processing this data locally without the need for an Internet connection, an iPhone can be a portable language translation device people can use anywhere in the world.
Machine learning is the future and Apple’s goal of creating a dedicated AI chip makes perfect sense to process this data rapidly on the device itself. That means tomorrow’s iPhone will not only include a general purpose ARM processor but also a graphics processor and an AI processor. Machine learning is going to be the future of mobile computing and Apple is leading the way.
Machine learning is simply the future of computing. It’s going to help widen the gap between Apple products and rivals to such an extent that it will be like comparing a word processor to a typewriter. Both will be able to type but which one will be more efficient and easier to use?
While Amazon and Google forge ahead with virtual assistants that gobble up market share, they’re missing one key element. Both Amazon and Google’s virtual assistants only understand a handful of languages. Siri understands far more languages and that means Siri will become the standard virtual assistant around most of the world.
The future is clear. Machine learning is what will separate tomorrow’s apps from today’s apps. The difference will be as vast as night and day.
To read more about machine learning, click here.