This blog answers the question – What is Machine Learning? And, it uses cats to answer. So, read on.
What is Machine Learning?
Most answers to the question ‘what is machine learning?’ either state the obvious and remain at the surface level – it’s the ability of machines to learn D’uh! Or, they go in too deep and leave you wishing machines never started to learn, which is not a very progressive view, I dare say.
So, here’s a way to think about machine learning.
Statement 1: Machine learning is effective categorization of things by computers.
I use the adjective effective because really computers are way better at this whole biz than us humans. So, let’s give our non-human friends some credit. In fact, they are so good at it that they can even categorize things that we don’t have a clue how to. That’s why it’s such a big deal.
How Do They Do It?
You see the image above? What do you see? I suspect cats. Because they are cats. How do you know? Well you probably learned to associate this set of adorable creatures with what the vast majority (or all) of English-speaking people call it – cats. And, that’s how you know they are cats.
But, how could a computer identify a cat or cats? This would be a problem that programmers worked on traditionally. They would probably stare blankly into the screen, zoom-in on the pixels, suffer from exhaustion and then when inspiration strikes, they would perhaps find a limited solution to the cat labelling problem. This, despite, the programmer being highly talented, mind you.
Teaching a computer or a machine to recognize cats might have been a problem in the past. Not anymore with machine learning, because computers are efficient like I said. All we need to do now is present a bunch of pictures of cats along with a bunch of pictures that have no cats and tell the genius computers to figure out a way to recognize cats. Easy right? Which brings us to the next aspect of machine learning.
Statement 2: Machine learning uses examples instead of instructions to categorize things.
By things, I don’t mean just images. Image recognition is only one use of machine learning. Machine learning is useful in classifying all sorts of things – videos for surveillance, emails to avoid spam, fraudulent activities from non-fraudulent ones for security, etc. Spotify uses machine learning to identify songs that are similar to each other and make music recommendations to over 100 million of its users.
Now, do you think you have the answer to the question – what is machine learning?
If you expected machine learning to be high-tech, then I probably disappointed you by saying its only a machine’s ability to categorize things. Machine learning hasn’t advanced to the stages we fantasize about in sci-films or novels. Nonetheless, the feats computers achieve in categorization are incredible. Even if it is not phenomenal in character, it is thoroughly useful. When we combine statements 1 and 2, we get the answer to the question ‘what is machine learning?’:
Machine learning is the way computers effectively categorize things by using examples instead of instructions.
Even if machine learning is not as advanced as we would like it to be, the signs indicate that it will be. Machines will learn more and they will give more to society as we progress to the age of artificial intelligence.
Many scholars and tech enthusiasts believe we are in the age of ‘weak’ artificial intelligence. But, the age of ‘strong’ artificial intelligence is surely upon us. And machine learning is drawing it ever closer.
I hope this blog answered the question – what is machine learning. To learn about the machine learning skills that are in demand, check out my blog on the In-Demand Machine Learning Skills.
Machine learning is expected to have a huge impact on how societies function. For more updates on this exciting technology, make sure you subscribe to Acadgild.