PMR2728/5228 - Probability Theory in AI and Robotics

Escola Politecnica, USP

Fabio G. Cozman

Basics

The objective of the course is to present a self-contained description of probabilistic techniques that have had a significant impact on the fields of artificial intelligence and robotics. The course consists of 12 classes, each lasting 3 hours. The first two classes are mostly devoted to basic theory, with forays into history and interpretation. The remaining classes are generally divided into a first block with theory, a second block with applications, and a third block on advanced concepts.

The syllabus contains the sequence of topics discussed in class.

Support material

  1. Commented syllabus.

    Link to paper In defense of probability, by Peter Cheeseman.

  2. Historical review.
  3. Basics on probability theory (finite spaces, subjective perspective).

    Support text and exercises.

  4. Bayes nets and Markov random fields, Part I and Part II.

    Link to nice material on graphical models (support).

  5. Infinite spaces.
  6. Decision making.

    Related exercise set.

  7. Basics on machine learning and classification.
  8. Bayesian network learning.
  9. Markov Chain Monte Carlo methods.
  10. Sequential decision making.


Page updated by Fabio Cozman (21/10/2013).