Week 3 Part I · Foundations
Multilayer perceptrons; the forward pass; backpropagation mechanics via PyTorch autograd.
Curated, free, canonical references for this week: a course or lecture, a book chapter, a video, and an authoritative blog post or official tutorial. Each opens in a new tab.
Builds an MLP and a tiny autograd engine from scratch, deriving the forward pass and backprop step by step.
Introduces hidden layers and nonlinear activations and why MLPs go beyond linear models.
A visual walkthrough of how gradients flow backward through a network to update weights.
Connects backprop theory to PyTorch autograd: the dynamic graph, requires_grad, and .backward().