The Contents
Each chapter is a plate of related ideas. Open one at any depth — every section is written for the newcomer, the practitioner and the researcher.
Chapter I Level: Foundations
Foundations of Learning Probability · Linear algebra · Gradient descent · Linear regression
Explainers · Plates · Exercises
Chapter II Level: Foundations
Classical Machine Learning Decision trees · Boosting · SVMs · k-means · PCA
Explainers · scikit-learn · Data sheets
Chapter III Level: Practitioner
Deep Learning Neural networks · Backpropagation · Optimizers · Regularization
Plates · Derivations · PyTorch
Chapter IV Level: Practitioner
Computer Vision Convolutions · CNN architectures · Detection · Vision transformers
3D plates · PyTorch · Papers
Chapter V Level: Practitioner
Reinforcement Learning MDPs · Q-learning · Policy gradients · RLHF
Plates · Simulations · Exercises
Chapter VI Level: Research
Generative Models Autoencoders · GANs · Diffusion · Language models
Derivations · PyTorch · Papers
Chapter VII Level: Research
Papers, Decoded Attention Is All You Need · ResNet · DQN · and more
The idea · Prerequisites · Interview
Not sure where to begin? Follow a path through the atlas → Guided routes for the newcomer, the practitioner and the researcher. Explainers Each idea written three ways, with moving plates you can turn and test.
Libraries scikit-learn and PyTorch, every argument tied to the idea it controls.
Papers, decoded The one idea worth keeping from landmark research, and the questions it raises.
Exercises Predict, drag, derive and debug — checked as you go.