01
Neural networks and backpropagation
How does a network know which weight to blame?
CourseFree
Neural Networks: Zero to Hero ↗Andrej Karpathy
You build backprop, then a transformer, from scratch, live. Free.
AI · Module 13
Deep Learning
Why a stack of matrix multiplications and one nonlinearity turned out to be enough.
By the end of this module
You can build, train and debug a neural network, and explain why depth helps.
0%0 / 3 complete
02
Optimisation, initialisation and regularisation in practice
Why is training a network more craft than theory?
InteractiveFree
Dive into Deep Learning ↗Zhang, Lipton, Li & Smola
Free interactive textbook with runnable code in every section.
03
CNNs, RNNs and why attention replaced them
What inductive bias does each architecture encode?
CourseFree
CS231n: Deep Learning for Computer Vision ↗Stanford University
Notes and assignments, free.
Quiz
Not written yet for this module. The lectures above are complete and the module still counts toward your progress.