Table of Contents

Fundamental Math Courses

All topics in one resource

• NYU DS-GA 1002: Statistical and Mathematical Methods - Breadth Course covering Probability, linear algebra, statistics, and optimization. Recommend doing each individually below
• Mathematics for Machine Learning by Marc Peter Deisenroth (Assistant Professor at Imperial College London)

Individual Classes/Books per resource

Probability

• pre-requirements: None
• Online Courses
  • Harvard - Statistics 110: Probability
  • edX MIT Course - Introduction to Probability - The Science of Uncertainty - RECOMMENDED
  • MIT OCW Course: Probabilistic Systems Analysis and Applied Probability
  • EXCELLENT THEORETICAL YET PRACTICAL SERIES by Michael Betancourt
    • Probability Theory (For Scientists and Engineers)
    • Conditional Probability Theory (For Scientists and Engineers)
    • Probabilistic Building Blocks
    • Probabilistic Modeling and Statistical Inference
• Books: 
  • A First Course in Probability by Sheldon Ross
  • Monte Carlo theory, methods and examples (last 2 chapters on Variance reduction and Importance sampling are particularly useful in machine learning)

Linear Algebra

• pre-requirements: None
• Online Course: 
  • MIT OCW Course: Linear Algebra (lectures, hws, exams, solutions available) - RECOMMENDED
  • Introduction to Applied Linear Algebra – Vectors, Matrices, and Least Squares - Course by Stephen Boyd
• Books/PDFs:
  • Short: 11-page review by NYU computational neuroscientist Eero Simoncelli
  • Medium-Length, gentle introduction: 58-page Appendix on Linear Algebra from the PDP book series, by computer scientist Michael Jordan
  • Gilbert Strang: Introduction to Linear Algebra
  • Sheldon Axler: Linear Algebra Done Right
  • The Matrix Calculus You Need For Deep Learning

Statistics

• pre-requirements: Probability, Linear Algebra
• Recommended Book Sequence:
  • An Introduction to Statistical Learning, with Applications in R (more intuition, less math)
    • Online Class
  • The Elements of Statistical Learning: Data Mining, Inference, and Prediction (more rigorous)
• Online Courses
  • Stanford Statistics 200: Introduction to Statistical Inference (more rigorous, has solutions for psets/exams) - RECOMMENDED
  • MIT OCW Course - Statistics for Applications (more rigorous, lacks solutions for psets/exams) - proof based
    • con: doesn’t teach intuition for complex math. requires mathematical maturity.
  • Stanford: Statistical Learning (less rigorous, everything available online) - good beginner course. more intuition, less math.

Convex Optimization

• pre-requirements: Probability, Linear Algebra
• Convex Optimization Class, Lectures - RECOMMENDED
  • Courses: EE364A (Stanford) or EE236B (UCLA)