Deep Learning

Neural networks from the single perceptron to modern transformers: how depth, the right activations, and gradient-based training let a model learn its own features instead of hand-crafted ones.

Prerequisites: the Machine Learning course (especially the perceptron in Linear classification), basic Python, calculus, and linear algebra.

Syllabus

1. Introduction
2. Multilayer perceptron
3. Activation functions
4. Loss functions and output layers
5. Backpropagation
6. Optimization
7. Initialization and vanishing gradients
8. Normalization
9. Regularization and dropout
10. Convolutional networks
11. CNN architectures
12. Embeddings and representation learning
13. Recurrent networks
14. LSTM and GRU
15. Attention
16. Transformers
17. Deep learning in practice

Machine Learning · MLOps · Home