Ask most people what the phrase ‘machine learning’ means to them, and they’ll tell you it conjures up images from every science fiction movie nightmare where robots wiped out humanity. Movies such as The Terminator or The Matrix seem to serve as warnings against allowing AI to grow too smart. That’s really not the case […]
Ask most people what the phrase ‘machine learning’ means to them, and they’ll tell you it conjures up images from every science fiction movie nightmare where robots wiped out humanity. Movies such as The Terminator or The Matrix seem to serve as warnings against allowing AI to grow too smart.
That’s really not the case here; ‘machine learning’ is the term that describes the act of data sharing between machines, allowing devices to learn, and make helpful predictions based on a user’s experience of a specific device.
In other words, your smartphone, thanks to machine learning, is geared towards giving you the most personalised experience possible, regardless of its initial programming, as well as improving things the more you use it.
The iPhone’s virtual assistant Siri is a great example of this. In previous iterations of the iPhone it was taking baby steps. Now thanks to machine learning it can understand semantics and nuances in languages and tones. It even understands sarcasm – seriously, ask Siri what 0 divided by 0 is.
Where once Siri was a search engine requiring verbal input, on the iPhone 8, 8 Plus and iPhone X, now it’s an AI in your hand that can improve and optimise your experience the more you interact with it.
There’s a reason the three new Apple handsets have been called “the future of the smartphone” by some. Each one of them boasts the A11 bionic chip under the hood – a ‘neural engine’ – that is geared towards improving on machine learning algorithms. Apple says that the chip processes “up to 600 billion operations per second” to help the local AI learn from its user.
In other words, those interactive facial emojis touted in the iPhone X reveal just may have a bit more going on behind the scenes. As has your fitness progress on an app like Strava or your taste in music, films or videogames. However, with the new iPhone range there’s a rather welcome twist; whereas in the past the data the local AI used for machine learning was housed in a cloud server, now it sits directly on the device.
Your personalised iPhone experience just became a lot more personal.
AI goes to work
Machine learning, however, has myriad applications outside of smartphones – but don’t worry, it’s not going to replace human input anytime soon. The thrust behind the technology is aimed at taking the tedium out of everyday tasks.
As machines gain access to more and more data they can crunch and interpret, they free up their human counterparts to spend more time using their knowledge to analyse and interpret data, using time that would be otherwise occupied by their AI colleagues.
Google’s ‘DeepMind’ project is probably the best-known machine-learning endeavour – although up until now all this AI has been allowed to do is play games. Up until a couple of years ago, DeepMind was focused on the rather deviously complex board game called ‘Go’. After competing with some of the world’s best, two years later, DeepMind established itself as the world’s best player of this game.
Now, Google has partnered with game developer Blizzard (the company behind World Of Warcraft) in order to take DeepMind to the next level by playing the real-time-strategy game Starcraft 2. The pair have also opened this platform to approved developers for input as the AI goes about tackling the world’s most popular eSports game.
If this all sounds a little scary, don’t worry; even though DeepMind is at the bleeding edge of AI and machine learning, there are humans to keep things in check. And besides, ultimately, machine learning isn’t about creating a race of AI-controlled robot overlords. It’s really about using your everyday interaction with your smartphone or tablet to build the best experience possible for you.
As AI improves through machine learning, the smart devices we all use may learn more about us, but that’s ultimately to our benefit.
Most folk who own a smartphone have a computer in their pocket – one more powerful than a desktop was six years ago and one more powerful than a super-computer around two decades ago.
You shouldn’t be surprised it’s learning from you. And don’t worry, it has no affiliation with Skynet.