Subject

Probabilistic methods for reasoning and decision-making under uncertainty. Inference and learning in probabilistic graphical models; prediction and planning in Markov decision processes; applications to computer vision, robotics, speech recognition, natural language processing, and information retrieval.

Prerequisites

The course is aimed at advanced undergraduates and beginning graduate students in mathematics, science, and engineering. Prerequisites are multivariable calculus, linear algebra, elementary probability, and basic programming ability in some high-level language such as C, Java, or Matlab. (Programming assignments are completed in the language of the student's choice.)

Texts

• S. Russell and P. Norvig, Artificial Intelligence: A Modern Approach, Prentice Hall, 2003.
• R. Sutton and A. Barto, Reinforcement Learning: An Introduction, MIT Press, 1998.
• C. Bishop, Pattern Recognition and Machine Learning, Springer, 2006.
• T. Hastie, R. Tibshirani, and J. Friedman, Elements of Statistical Learning, Springer-Verlag, 2001.

Instructors

• Professor: Lawrence Saul
• Teaching assistant: Tingfan Wu

Meetings

• Lectures: Mon/Wed 12:30-1:50, EBU3B 2154.
• Sections: Fri 4:00-5:00, EBU3B 2109.

Grading

• homework assignments (50%)
• final exam (50%)

Assignments

• Assignment 1 - due Mon Oct 08
• Assignment 2 - due Mon Oct 15
• Assignment 3 - due Wed Oct 31
• Assignment 4 - due Wed Nov 07 [data]
• Assignment 5 - due Mon Nov 19 [data]
• Assignment 6 - due Wed Nov 28 [data]
• Assignment 7 - due Dec 07 [data]

Syllabus

• Mon Oct 01 - Administrivia and course overview.
• Wed Oct 03 - Modeling uncertainty, review of probability, explaining away.
• Mon Oct 08 - Belief networks: from probabilities to graphs.
• Wed Oct 10 - Conditional independence, d-separation, polytrees.
• Mon Oct 15 - Algorithms for exact and approximate inference.
• Wed Oct 17 - Learning via maximum likelihood estimation.
• Mon Oct 22 - CLASS CANCELLED DUE TO WILD FIRES.
• Wed Oct 24 - CLASS CANCELLED DUE TO WILD FIRES.
• Mon Oct 29 - Markov models of language, naive Bayes for text classification, linear regression.
• Wed Oct 31 - Logistic regression, gradient descent, Newton's method.
• Mon Nov 05 - Latent variable models, EM algorithm, statistical language modeling.
• Wed Nov 07 - Hidden Markov models (HMMs), automatic speech recognition.
• Mon Nov 12 - Veterans Day Holiday.
• Wed Nov 14 - Viterbi and forward-backward algorithms in HMMs.
• Mon Nov 19 - Multivariate Gaussian distribution, mixture models, Kalman filtering.
• Wed Nov 21 - Reinforcement learning, Markov decision processes (MDPs).
• Mov Nov 26 - Policy evaluation, policy improvment, policy iteration.
• Wed Nov 28 - Value iteration, stochastic approximation theory.
• Mon Dec 03 - Temporal difference learning, Q-learning, applications and extensions.
• Wed Dec 05 - Course wrap-up and evaluations.
• Thu Dec 13 - Final exam