“You do not really understand something unless you can explain it to your grandmother”

Not sure where this quote originally came from, it is sometimes kind of half-attributed to Albert Einstein.

Anyway, this post is my attempt of explaining (to myself and others) how neural networks (NN) algorithm works in a simple, novice, straightforward, and high-level fashion, without any formulas, equations, or codes. I am fully aware that there might be some inaccuracies in the text below, but this is naturally to happen in order to avoid complicated explanations and to just keep things simple, otherwise your grandma won’t understand…

Before we begin, I would like to acknowledge fast.ai, founded by Jeremy Howard and Rachel Thomas, which is a non-profit research group focused on deep learning and artificial intelligence. I have learnt and still learning a lot from their free online courses.

Oh… and I wrote a summary of this post at the end in case you want to skip to it.

OK, I assume that if you read this post you have at least some familiarity with NN, so I’ll save you the introduction about how NN became popular in recent years and how they can help you solve problems.

So how does NN algorithm works?

In one word: math.

In two words: matrix multiplication.

If your grandma did not understand this explanation then you can tell her that we start with an input layer, which is the input data for a NN model.

#neural-networks #machine-learning #deep-learning #deep learning