Machine learning
Neural networks and backpropagation
Layers, activations, losses, gradient descent and backpropagation, and why training goes wrong.
About 25 hours of study.
Where to learn it
Read
Dive into Deep Learning
Zhang, Lipton, Li, Smola
Chapters 3–6: linear networks, multilayer perceptrons, the builders' guide
Watch
Neural networks
Grant Sanderson (3Blue1Brown)
The whole series
The clearest picture of what backpropagation is actually doing.
Practise
Understanding Deep Learning notebooks
Simon J. D. Prince
Notebooks for chapters 3–9
Sixty-eight Python notebooks with the key code left for you to fill in.
Other options
- Neural Networks: Zero to Hero (watch)
The first lecture, building micrograd
Wrong chapter, dead link, or know something better? Tell us on GitHub.
Which roles ask for it
- Deep learning engineer (vision / language) to get hired
- GATE data science and AI to get hired
- UGC NET computer science to get hired
Before and after
Learn first
Neural networks and backpropagation