What is machine learning?
Rise of the robots…
03 Nov 2017
From science fiction to reality
Not so long ago, the idea of controlling gadgets using our voice seemed like nothing more than a science fiction fantasy. Now thanks to voice assistants such as: Apple’s Siri, Amazon’s Alexa, and Google’s Assistant, we’re all doing it.
Voice recognition is taking over, but it wasn’t always like this. The technology has come a long way in a short space of time. It’s changing the way we all use our smartphones, and home appliances. And it’s all thanks to machine learning.
What is machine learning?
Machine learning has its roots in the 1950s, when scientists taught a computer to play draughts. These computers learned to play like a human, by adapting to new patterns of play, and predicting moves before they happen.
To do this, a machine-learning algorithm was provided with data, a list of moves to play, and their possible outcomes. It used this list to learn how to play. With practice, the computer would learn to become a better player.
More famously was the Turing Test, devised in 1950 by WWII code-breaker, Alan Turing. Its aim was to test whether a machine could mimic natural human behaviour, indistinguishable from that of a person.
A human participant would read text responses by a man, and a machine. And was then tasked with the challenge of determining which responses were human, and which machine. Turing concluded in his studies that machines can learn to think like a human. And if the machine fooled the human into thinking its response weren’t artificial – it passed the Turing Test.
Now here’s where the likes of Alexa, Siri, and Google Assistant come in. Right now, any of these voice assistants can understand your voice commands. Ask a basic question, and you’ll get an answer.
But in time, these too will learn to become better, answer more complex questions, and be able to distinguish between different voices thanks to their own machine learning algorithms. They’ll be able to understand us better, listen to the way we speak, and decipher complex human language to answer any question we ask of them.
They’ll even sound more like us. Google Assistant, for example, uses a technology called WaveNet, which makes its speech sound more natural, and less robotic.
The future of smartphones
In a nutshell, this means you’ll soon be able to talk to Siri, Alexa, or Google Assistant as if they were real people, edging us closer to a world with menu-free, button-free smartphones and gadgets, purely activated and controlled by voice.
At a Google Developer Days conference earlier this year, Google demonstrated how its Google Assistant, on an Android phone, will be able to respond to normal everyday language.
In the video below, a Google spokesperson asks: “What is the name of the movie where Tom Cruise acts in it and he plays pool, and while he plays pool he dances.”
Amazingly, Google Assistant is able to find the answer, bringing up results for Cruise film, The Colour of Money. See it for yourself here:
In the same conference, Google also demonstrated how its Assistant can learn to better understand the context of questions being asked. The presenter asked for photos of Thomas. With no context, the Assistant pulled up pictures of Thomas the Tank Engine.
It was then asked for Bayern Munich’s football team, which features a well-known player called Thomas Müller. The Assistant was once again asked for pictures of Thomas. This time it retrieved pictures of the footballer, demonstrating a contextual awareness previously only possible by humans.
This is just two examples of how machine learning will change the way we use our smartphones and household gadgets within the coming year alone. And it’s a huge step closer towards an inevitable future of technology controlled purely by voice.
Don’t worry if you feel uncomfortable at the thought of talking to a machine. You’re not alone. But write off voice control at your peril. It’s here to stay, all thanks to machine learning.