MACHINE LEARNING is the process by which a machine learns something. The end.
Only joking. We’re going to dig a little deeper than that, but it does go to show how simple the basic concepts of machine learning can be. In this article, we’re going to make machine learning so easy that a child could do it. That’s why we’re going to use LEGO.
Our machine learning example is going to identify information about each of the LEGO bricks including color, size and surface area. By storing information on each of those bricks inside the algorithm’s database, it can start to predict which pieces you might need next. In fact, it can start to analyze every possible combination of pieces and to identify shapes that you might be trying to build. Think of it as like Google’s autocomplete, but with LEGO.
Some machine learning algorithms are more flexible than others, but this flexibility usually comes with some other cost such as the amount of resources that are required to run the algorithm. On top of that, most developers will run different iterations of different algorithms to see which gets the best results.
We usually call these different iterations “models”, and it’s these models that we then tap into over time. With our LEGO blocks, the model would be able to classify new blocks as they’re added, and not just the blocks it was originally given. In fact, the true power of machine learning comes from the fact that the more it’s used and the more data that it processes, the better it gets.