Lecture Slides

Interactive Reveal.js presentations covering all ten parts of the course. Use arrow keys to navigate, S for speaker notes, ESC for slide overview.

Parts I & II

Origins & The Perceptron

Chapters 1–7 · 48 slides

McCulloch-Pitts neuron · Boolean logic · Perceptron model · Learning algorithm · Convergence theorem · Boolean functions

Part III

Limitations & Breakthroughs

Chapters 8–11 · 21 slides

XOR problem · Minsky-Papert · Linear separability · Cover's theorem · Multi-layer solution

Part IV

Learning Rules

Chapters 12–14 · 18 slides

Hebbian learning · Oja's rule · PCA connection · BCM rule · Credit assignment problem

Part V

Backpropagation

Chapters 15–19 · 32 slides

Gradient descent · BP derivation · Activation functions · Vanishing gradient · Universal approximation

Part VI

Synthesis

Chapter 20 · 13 slides

Grand timeline · Three themes · AI Winter lessons · What comes next

Part VII

Convolutional Neural Networks

Chapters 21–25 · 25 slides

CNN motivation · Convolution operation · Architecture from scratch · Training & filter evolution · Experiments & analysis

Part VIII

Modern Optimization

Chapters 26–28 · 23 slides

Cross-entropy & KL divergence · Adam optimizer · Automatic differentiation · Micrograd engine

Part IX

Introduction to PyTorch

Chapters 29–31 · 20 slides

Tensors & autograd · nn.Module · Training loops · MNIST MLP · PyTorch CNN

Part X

Recurrent Neural Networks & LSTM

Chapters 32–36 · 30 slides

RNN motivation · BPTT · Vanishing gradients · LSTM gates · Char-RNN on Shakespeare · Seq2Seq

Part XI

Attention & Transformers

Chapters 37–40 · 58 slides

Bahdanau attention · Scaled dot-product · Self-attention · Schmidhuber 1991 · Vaswani 2017 Transformer

Keyboard shortcuts: Arrow keys to navigate · S speaker notes · ESC overview · F fullscreen · ? all shortcuts

Total: 288 slides across 11 presentations covering all 40 chapters.