With an understanding of a single neuron (from the logistic regression example), we can now study networks of neurons, both shallow (e.g., with one hidden layer) and deep.
Topics to Look For¶
- Neural network structure
- Neural network computation
- Activation functions
- Random initialization
- Vectorized implementation
Resources¶
- DeepLearning.ai: Neural Networks and Deep Learning, Week 3, all videos (you can skip the one marked “optional”) (11 videos)
- DeepLearning.ai: Neural Networks and Deep Learning, Week 4, all videos (8 videos)
Supplemental¶
- In the 3Blue1Brown playlist of videos on neural networks, Chapter 3 will give another perspective on backpropagation, while Chapter 4 will delve into the mathematics more, if you’re interested.
- Chapter 1 from Neural Networks and Deep Learning covers most of these topics in some depth.
- Chapter 6 from the Deep Learning book covers them as well, in a fairly math-heavy fashion.