Stanford CS229: Machine Learning

Andrew Ng's foundational machine learning course. The gold standard for ML education.

About This Course

Stanford CS229 is the course that launched the modern machine learning education movement. Taught by Andrew Ng (and more recently by other Stanford faculty), it provides a rigorous mathematical foundation for understanding machine learning algorithms.

The course covers both the theory and practical application of ML, with an emphasis on understanding why algorithms work rather than treating them as black boxes.

What You Will Learn

  • Supervised Learning: Linear regression, logistic regression, generalized linear models
  • Optimization: Gradient descent, stochastic gradient descent, Newton's method
  • Generalization: Bias-variance tradeoff, regularization, model selection
  • Neural Networks: Backpropagation, architectures, training techniques
  • Support Vector Machines: Kernels, duality, the kernel trick
  • Unsupervised Learning: K-means, EM algorithm, PCA, ICA
  • Reinforcement Learning: MDPs, value iteration, policy iteration, Q-learning
  • Practical Advice: Feature engineering, debugging ML algorithms, system design

Prerequisites

Linear algebra, probability and statistics, multivariable calculus. Programming in Python/NumPy.

Course content belongs to Stanford University.

Up next

MIT 6.034: Artificial Intelligence

Patrick Winston's classic AI course. Covers search, constraint propagation, learning, neural nets, genetic algorithms, and support vector machines. A historical gem.