The 3 Facts You Should Know About Machine Learning

Those of you connected to the tech world day and night may have noticed the growing trend in some news: “Scientists use machine learning to…”, “A program just surpassed humans at… thanks to machine learning”, and other such eye grabbing articles.

Indeed, machine learning, and AI, are rising at a spectacular rate. Makes one wonder, exactly what is machine learning? What can it do? How will it impact the future?

I’m gonna try to boil down machine learning into 3 manageable chunks, to put you into the 5% of people who understand this amazing new technology, and what it’ll mean for the future! Lets get started.

1. It’s a method of analyzing data – that teaches computers how to LEARN

Here’s a little fact you may not fully understand about your PC - it cannot truly think. Everything your PC does (like any other machine), is take what you put in, and spit out what you want it to, always using the same methods.

Although it can do things you would never be able to do in your life, it has a flaw – unlike you, it cannot learn. After doing task A two hundred times, it will continue to do it as it always did. Whereas you, if trained to, would gradually get better at any task through experience, and learning.

How exactly do you do that? Simple. Lets make an analogy: think of when you started typing on a keyboard.

At first, you didn’t have any data on how it worked – you needed to look at every key as you typed. As you kept doing it though, something amazing happened. Your brains had registered the position of every key, so now, your brain had a map of what finger should go to what place on the keyboard to hit that key. The process continued, till the point where you can now very likely write without even looking at the keyboard.

What your brain did was make a model of the keyboard, and used it everytime you had to write, in order for you to make better time. Everytime you used it, the model grew better, gradually improving your skill at typing. Learning is all about building efficient models of information and honing them through practice.

That, dear reader, is what machine learning allows a machine to do. While doing a task, it can collect data about that task, and for the first time in the history of computers, manipulate it, make a model of it, so it can do it again much faster, better, and cheaper next time.

2. It can be tailored to suit almost ANY TASK

The breadth of potential this technology has is astounding. As it turns out, there are a lot of tasks a computer can do better given enough practice.

For instance, these systems can be used to analyze a ton of stock data, and predict whether the prices will go up or down. These new systems can shotcall better than humans, and can be used to create an almost risk-free portfolio, if fed enough data.

If given enough data about a user, a machine learning system could progressively give the user shopping suggestions, performing better and better each time…

They can also detect fraud (caused by security data leakage) far better than any human could. Government and financial institutions alike have turned to them for such purposes.

Also, they could be used to create biometric data systems, which have the potential to erase the need for passwords and usernames. Remember the clunky password you were given upon launching steemit? That could soon be a thing of the past…

After covering so many uses, you know the best part? I could easily list even more. These systems are only limited by the creator’s imagination…

3. It’s already WIDELY used

You wouldn’t believe how much these systems have already been integrated in our daily lives, without us even knowing. You don’t even need to go into a highly advanced top-secret government lab to see these systems in action. You only need to log into your email.

See your spam folder? In case you use Gmail, you should know that that spam folder works through machine learning. This nice folder has been trained with a lot of data, and is able to separate the good from the bad, making your email experience much easier.

Another surprising way to find these machines in action? Click the “Contact Us” part of any major website. If it keeps its gear up to date, this section will have machine learning algorithms that will simply allow the user to describe a problem generically, then send them out to the right people to handle it within the company, instead of you having to fill numerous forms.

My own antimalware service, AntiMalwareBytes 3.0 (fully recommended) warned me that it would now use machine learning mechanisms in the latest update, to better search for viruses in my pc.

Depending on where you live, your local hospital may even be using a machine learning algorithm to check out which patients are more likely to be doing well within the next 5 years, and which will most likely need care, allowing them to prepare for such eventualities.

Amazing stuff, right?

It is indeed. This technology has slowly crept up on us these last few years, and it’s here to stay. As it develops, its amazing abilities will only increase.

I should warn you though – this is not the same as AI. I’ll be covering that in a later article. And you should, by all means, investigate this technology. You’ll soon be having it in many sides of your life anyway!

If you liked this article, and would like to see more content like this, feel free to follow, and leave a comment/question! If you have any tips to help me improve, I wanna hear them! :D