Neural Networks

1. Perceptron and MLP

   Start here

   Write a forward pass for a small MLP, and say what a nonlinearity is for.

   Take this classAlready know this

2. Backprop

   After Perceptron and MLP — test out to unlock

   Compute gradients for a tiny net by hand, and match them to what autodiff would give.

   Test out to unlock

3. Convnets

   After Backprop — test out to unlock

   Explain convolution, pooling, and a parameter count on a small image model.

   Test out to unlock

4. Sequence models

   After Convnets — test out to unlock

   Contrast an RNN step with attention at a cartoon level, and say what each is remembering.

   Test out to unlock

5. Optimisation

   After Sequence models — test out to unlock

   Use SGD vs Adam, learning-rate, and batch size, and diagnose exploding/vanishing loss.

   Test out to unlock

6. Generalisation

   After Optimisation — test out to unlock

   Use dropout, augmentation, or early stopping, and read a train/val gap.

   Test out to unlock

Why a path, not a playlist

MLPs, backprop, convnets, optimisation. Later headings open after the one before.

Each heading is its own course with a tutor. Later headings open when you finish — or test out of — the one before.

After you sign in

• •You land on this map — every heading is already there.
• •Take the open class, or test out of a later heading to unlock it.
• •Each heading is one shared course — same syllabus and first class for everyone. Later classes continue on your copy.