Handouts and course notes
- Handout 0 - about the course
- Lecture 0 - Introduction to the Statistical Learning Sequence
See also this short presentation
- Lecture 1 - Basics of Multivariate Inference (8 pages)
- Lecture 2 - Independence and conditional independence (5 pages)
- Lecture 3 - Graphical representations of conditional independence. Part I - Markov Random Fields (10 pages)
- Lecture 4 - Graphical representations of conditional independence. Part II - Bayesian Networks (11 pages)
- Slides on Gaussian Graphical Models by Steffen Lauritzen
- Lecture 5 - Decomposable graphical models, triangulation and the juntcion tree (9 pages)
- Lecture 5.1 - Juntcion tree - additional proofs (2 pages)
- Lecture 6 - Variable elimination (8 pages)
- Lecture 7 - The Junction Tree Algorithm (10 pages)
- Lecture 8 - The Junction Tree Algorithm: Remarks, Variants, Sum-Product Algorithm (8 pages)
- Factor Graph
s and the Sum-Product Algorithm by Kschischang, Frey, Loelinger, 22 pages -- the sum-product algorithm for l oopy graphs; factor graph transformations (like clustering and stretching); a FFT algorithm from factor graph transformations; other algorithms from the point of view of factor graphs
- Lecture 9 -- not posted
- Lecture 10 - Max Propagation with The Junction Tree Algorithm (6 pages)
- Lecture 11 Estimation of graphical models parameters (13 pages)
- Lecture 12 - Iterative Proportional Fitting (5 pages)
- Lecture 13 - Structure learning notes (to be posted)
- Lecture 14 - Classic and modern data clustering (SLIDES: 131 pages)
Work sheets
|