CS 501
Fall 2026
Graduate
Machine Learning Foundations
Core principles of statistical learning, optimization, and neural networks with an emphasis on rigorous evaluation and reproducibility.
Course overview
This course covers the mathematical and practical foundations of modern machine learning: statistical learning theory, convex and non-convex optimization, and the building blocks of neural networks. The course emphasizes reproducible experimentation — every assignment requires a written report alongside working code.
Topics covered
- Supervised learning: linear/logistic regression, regularization, generalization bounds
- Optimization: gradient descent variants, convexity, convergence guarantees
- Neural networks: backpropagation, initialization, normalization
- Model evaluation: cross-validation, statistical significance, common pitfalls
- Introduction to fairness and robustness considerations in learned models
Prerequisites
Linear algebra, multivariable calculus, and a probability course. Prior programming experience in Python is expected; no prior machine learning coursework is required.
Grading
- 40% — Four problem sets (theory + implementation)
- 30% — Final project (proposal, implementation, written report)
- 20% — Midterm exam
- 10% — Participation and in-class reproduction exercises