Part 1: What is Machine Learning?
The quest for Artificial Intelligence is a field which has both fascinated and terrified mankind for decades. Books like Isaac Asimov’s “Robot” series and Douglas Adams’ “Hitchiker’s Guide to the Galaxy” envision dystopian futures in which Intelligent machines have become indispensable to human life, whilst films like “The Terminator” and “Ex Machina” explore a darker side of AI, where machines become a threat to mankind’s very survival.
So what is the truth? Do we get C-3PO, or do we get Ultron?
Thankfully, that is a debate for another generation to grapple with. For the moment, things are much simpler. At this point, AI- a machine which mimics the human mind, is still a pipe dream.
Machine Learning, however, has been a reality in our lives for quite some time.
“Machine Learning” has been defined thusly:
A computer program is said to learn from experience “E” with respect to some class of tasks “T” and performance measure “P” if its performance in tasks “T”, as measured by “P” improves with experience “E”
This, of course, is just a fancy way of saying that if a machine is able to perform a task more effectively over time based on measuring its own performance and changing how it performs its tasks accordingly, it can be considered a learning machine.
Interesting stuff, but what do we do with it?
Today, as ever, mankind has put most of its collective resources in this area into finding ways to get these learning machines to make us money.
And make us money they have.
How, you ask?
It all starts with algorithms.
It turns out the world is not as chaotic and random as we once thought. Almost everything is predictable on some level, even human behaviour. If one is able to effectively recognize patterns, and use these patterns to anticipate future events, one can do a lot of things much more effectively.
This is what algorithms do.
We won’t bore you with a bunch of math here, you’ll just have to take our word for it for the moment. Mining and compiling enough data and exhaustively analyzing all the variables involved may not produce perfect predictions of future events, but it can get you pretty darn close.
Just ask any bookie.
Do you really think Knuckles here would be able to afford a fabulous fedora like that if he didn’t know something that all the punters placing bets with him didn’t?
And, as a follow up question, does this fellow look like the type to spend countless hours poring over sports statistics, furiously pounding his chubby fingers at the buttons of a calculator?
Probably not. In this scenario, Knuckles is the brawn of the organization. The brains look a little bit more like this:
Machine learning in it’s purest form.
Luckily, Knuckles doesn’t need to know the exact score of tonight’s game, which is good, because we don’t have computer that can predict that yet.
What we do have are algorithms which can measure an extensive variety of variables.
How does team x typically perform under certain weather conditions? How many points does team Y typically score when player A is out with an injury? How well does team Z defend its goal in games played after 7:00pm?
The variables are endless, and they are simply too much for the human brain to effectively analyze. But a machine can. And that, dear reader, is Knuckles’ edge. He doesn’t need to know everything, he just needs to know more than you.
That is why he is driving a $60,000 car, while his customers are riding the bus with broken fingers.
So where does the “learning” part come in to play?
When it comes to making money, enough is never enough. For a business, there are always new markets, new products, new market trends, new competitors, and myriad other drivers forcing it to change and adapt its strategy to market its goods and services more effectively.
A business needs to keep on growing, and this means adjusting.
In the old days, this meant board meetings, endless strategic planning sessions, task forces, and think tanks. Today, with machine learning involved, the process happens in real time, with little or no interruption to the business day.
The machines involved learn as they go. If a process or strategy is not producing good results, it is eliminated, If a process or strategy is excelling, it is pushed to the forefront, honed, and perfected. This is the beauty of the machine learning process.
This, in a nutshell, is how machine learning can benefit any business. Every business has tasks which need to be completed. The efficient completion of these tasks allows a company to become more profitable, reach more consumers with its advertising, sell more, and cut down on waste.
As far as how this process can be specifically applied to businesses, we will get into that in part 2 of this series;