Syllabus Lesson 2 of 246 · Module 0: Start line
Module 0: Start line

Bridge check: what you already know

Why it matters This course starts from one neuron, but it moves quickly. This short check tells you whether to read the quick review below before going on.

Five questions on the basic ideas. If you miss more than one, read the review first; it takes about 20 minutes.

Quick review

A neuron multiplies each input by a weight, adds them up, adds a bias, and passes the total through an activation function. A layer is many neurons side by side, each with its own weights. A network is layers in a row: each layer's outputs are the next layer's inputs.

Training means finding weights that make the loss small. The gradient of the loss tells you, for each weight, how the loss changes if that weight is nudged. Gradient descent nudges every weight a little downhill, again and again. Backpropagation is the efficient way to get all those gradients at once, by passing sensitivities backward through the calculation. You will build every one of these from scratch in Modules 2 to 4, so a rough picture is enough for now.

Words and symbols in this lesson
loss
one number saying how wrong the model is; training makes it smaller

Spotted a problem in this lesson? Report it · Ask about this lesson

This exercise opens soon

Finish the lesson before it and this one unlocks. The whole course is free, no account needed, and you can read this lesson right now.