Deep Learning · MT2 · guide

Neural-network foundations

1. What does learning from data mean
2. Your first complete training step
3. Classification regression and learning types
4. Training validation and test data
5. Parameters and hyperparameters
6. What one artificial neuron does
7. Model answer: explain an artificial neuron
8. Activation functions and how to calculate them
9. Model answer: why a network needs nonlinear activation
10. Choosing a loss
11. Forward propagation through multiple neurons
12. Backpropagation with one hidden neuron
13. More outputs and vectorized backpropagation
14. Model answer: forward propagation, backpropagation and updating
15. Counting parameters in a dense network
16. What you should understand before optimization
17. Complete sigmoid training example
18. Cross-entropy for probability distributions