In no specific order:

1. Gibbs Sampling
2. Parametric vs non-parametric models: While the “non-parametric” might suggest that the model has “no parameters”, they no fixed number of parameters, the parameters used by the model will simply grow with the number of training examples seen. Simple example of non-parametric model is the KNN classifier.
3. On-policy vs Off-policy, SARSA vs Q-Learning link
4. Value iteration in POMDPs section 12.3.2 in Reinforcement Learning: State of the Art
5. A3C tutorial