My study notes on statistical learning and reinforcement learning, with an emphasis on mathematical reasoning and the connections between methods.
Statistical Learning
Lecture notes from UW-Madison, organized into four chapters: prediction and consistency; empirical risk and generalization; linear classifiers and SVMs; and kernels, Gaussian processes, and conformal prediction.
Start with Prediction and consistency, or browse the course contents above.
Reinforcement Learning
Mathematical Foundations of Reinforcement Learning
Notes on Shiyu Zhao’s book, connecting Bellman equations, dynamic programming, Monte Carlo and temporal-difference learning, value approximation, policy gradients, and actor-critic methods.