Lesson 1: Is It A Bird?

This lesson was mostly a rough overview of the ideas that can be summarized as “What is deep learning?“. Even as recent as 2015, we considered writing a program that can identify a bird in an image as basically impossible.

Now that’s a trivial task, basically the “hello world” of deep learning.

The instructor demonstrates this using fastai — a library on top of PyTorch that provides a nice, easy-to-use, high-level API for beginners to deep learning. The instructor also demonstrates how cool and well-revered he is, how to use Jupyter notebooks for everything, and some of the basic concepts behind deep learning.

This is the very basic idea of machine learning. Thinking of a machine learning model as a normal computer program, but the weights affect the outcome of the program. We update the weights behind the model after running the program. Boom.