Statistical Learning Theory
Compute the bound next to the quantity it bounds.
-
-
Markov Decision Processes and Dynamic Programming
The classical half of sequential decision making, and the prerequisite for deep RL.
-
Bandits and Exploration
A bandit earns more reward and gives you a worse estimate of the effect.
-
The Bayesian Workflow
A distribution over parameters costs a prior, a sampler, and two diagnostics that can both be wrong.