Schedule

The following schedule is possibly optimistic, but is definitely tentative and subject to revision, depending on how we progress. Ideally, we'll touch on all of the topics given here (and maybe some additional), but we may dwell longer on some, in the interest of improved understanding, at the cost of neglecting others.

Weeks 1-2

Introduction; basic concepts and definitions; examples; empirical methodology.

Weeks 3-5

Introductory classification and regression; decision trees; classification by linear machines; support vector machines; linear regression.

Weeks 6-8

Introduction to generative classification; Bayes' rule; principles of Maximum Likelihood and MAP; graphical models.

Weeks 9-10

Unsupervised learning; mixture models; expectation-maximization; graphical models with hidden variables.

Week 11

Time series models; Markov chains; hidden Markov models.

Weeks 12-14

Reinforcement learning and planning; Markov decision processes; POMDPs; Q-learning; $E^{3}$, Reinforce.

Week 15

Final project presentations.

Final Exam Week

The final exam will be in ME218 on Tues, Dec 14 at 5:30-7:30 PM.