Before you can start translating German to Afrikaans, and predicting the Stock Market, you need to understand what Artificial Intelligence is, and how it works.
What is Deep Learning/AI?
AI is the concept of making machines intelligent – allowing them to think like a human, and have a “brain”. You may have heard of
Artificial Neural Networks, or ANNs. Deep Learning is meant to imitate the process of the brain, using a system of Neurons and Weights. We get into some complicated concepts, which, if you are interested in the math/science behind it, you can read about
here. In the most simple sense, AI allows a machine to receive a bunch of X variables and return a Y.
Unfortunately, an AI model can’t just “automatically” learn, per se. Instead, it works just like the human brain – it learns from experience. If I were to see a black cat at 8 am when I biked to school every morning, I would begin to expect to see that same cat every day at 8 am. In the same way, we have to feed a machine a lot of real-life examples in order for it to find patterns and predict what will happen in the future, for example. So we get a dataset – a giant “spreadsheet” of data. There are many deep learning datasets
available online, but all the “good” and “excitingly relevant” datasets such as Self-Driving Car data, etc., are mostly privately owned and kept hidden. Companies like Google (now
Waymo) and
Uber want to limit the amount of competition in the market space. They have enough money and Data Analysts that they can actually create their own datasets. We’ll be doing the same in this tutorial.
Another important thing to know are the basic concepts of Training and Testing. We feed the machine our part of our dataset for it to learn patterns between the variables we feed it – this is called training. Then using what the machine has learned, we can then test it by feeding it Input Values (X variables) from the rest of our unused dataset, and compare what it predicts (Y predictions) to the actual values (Y variables).
1 Response
[…] We’ve already trained our model – it should already know that Y is 1 if X3 is between the other two variables (X1 and X2), otherwise Y is 0. We did this in Part 6 of our series, training the model on Training Data. Now we are programming our algorithm to take X1, X2, X3, and, using a trained model, predict what Y is for other Test data. You can learn more about Training and Testing Data from our “What is Deep Learning?” post. […]