Machine Learning algorithms are everywhere.
They're helping you pick out your next gift on Amazon, controlling what you find on Google, they're suggesting new music for you on Spotify, and they're doing their best to keep you on their website.
When trying to explain how machines learn we tend to try and describe it in human terms. Unfortunately (or fortunately) machine learning isn't based on how we teach humans.
The video is a bit simple in its explanations, but it describes some important concepts. So, watch how machines learn.
The video focuses on Genetic Algorithms, which is one type of machine learning, and neglects some of the other more complicated approaches.
As machine learning gets more complicated and evolved, it gets harder for a human to understand what makes it good … and that's okay.
It's human nature to feel safer when we understand something. It's human nature to envision machines as making human-like decisions, just faster.
Of course, just because it suits human nature to believe something, that doesn't make it true.
Part of what makes machine learning exciting is that it can do a lot of things well that humans are really bad at.
In reality, it doesn't matter why a bot is making a decision, or what inputs the bot is making the decision on. What matters is the performance and level of decision-making in relation to itself and to other options.
One of the lessons I teach when I speak publically is that in trading, caring about the markets is a weakness.
It's a distraction.
I don't care how markets are doing.
I care how my systems are doing. I care how my portfolio is doing.
There's too much data for me to try and care about anything else.
It's a brave new world, and not only is big brother watching, but algorithms are too.
