- Instructor: Hesam Montazeri (hesam.montazeri at ut.ac.ir)
- Teaching Assistants: Naser Elmi (naser.elmi at ut.ac.ir) & Samaneh Maleknia (maleknias at ut.ac.ir) & Fahimeh Palizban (fahimehpalizban at ut.ac.ir)
- Time & Location: Feb-July 2019, lectures are held on Wednesdays 10:00 to 12:00 at Ghods st. 37, Department of Bioinformatics, IBB, Tehran.
- Probabilistic Graphical Models by Daphne Koller & Nir Friedman.
- Pattern Recognition and Machine Learning by Christopher Bishop.
| Date | Lecture | Reading Assignments | Homeworks & Projects |
|---|---|---|---|
| 16/11/1397 | Lecture 1- Introduction to probabilistic graphical models (slides) | Required: Koller Textbook, Sec. 1.2, 2.1 Optional: Koller Textbook, Sec. 2.2 |
Project 1 Project1.R HW1 |
| 29/11/1397 | Lecture 2- Bayesian network representation (slides) | Required: Koller Textbook, Sec. 3.1, 3.2 | HW2 |
| 6/12/1397 | Lecture 3- Bayesian network representation 2 (slides) Tutorial 1- Introduction to R by Samaneh Maleknia (slides, R Script) |
Required: Koller Textbook, Sec. 3.3.1, 3.3.4, 3.4.1, 3.4.2 Optional: Koller Textbook, Sec. 3.3.2, 3.3.3 |
HW3 |
| 13/12/1397 | Lecture 4- Conditional probability distributions; Gaussian Bayesian networks (slides) Tutorial 2- Lagrange multiplier (notes); multivariate Gaussian distributions (slides) |
Required: Koller Textbook, Sec. 5.1-3, 5.4.2, 5.5, 7.1-2 Optional: McGeachie, 2014 (CGBayesNets: Conditional Gaussian Bayesian Networks) |
HW4 HW4.R HW4train.csv HW4test.csv |
| 20/12/1397 | Lecture 5- Learning Bayesian networks (MLE, fully observed data) (slides) Tutorial 3- Introduction to overfitting, Beta and Dirichlet distributions (slides) |
Required: Koller Textbook, Sec. 17.1-2 Optional: Koller Textbook, Ch. 16 |
HW5 HW5.csv |
| 21/01/1398 | Lecture 6- Bayesian paramater estimation (slides) | Required: Koller Textbook, Sec. 17.3 | Project 2 HW6 |
| 28/01/1398 | Lecture 7- Bayesian parameter estimation; constraint-based structure learning (slides) Tutorial 4- Introduction to information theory and JAGS (slides) |
Required: Koller Textbook, Sec. 17.4.1-3, 18.1-2, 3.4.3 | HW7 JAGS-Example |
| 04/02/1398 | Lecture 8- Score-based structure learning (slides) | Required: Koller Textbook, Sec. 18.3, 18.4.1 | HW8 |
| 11/02/1398 | Lecture 9- Score-based structure learning; partially observed data (slides) | Required: Koller Textbook, Sec. 18.4.1-3, 19.1.3, 19.2.1, A.5.1, A.5.2.1 Optional: Koller Textbook, Sec. 18.5 |
HW9 |
| 18/02/1398 | Lecture 10-Expectation-maximization; regulatory motif finding (slides) | Required: Koller Textbook, Sec. 19.2.2 and Section 2.1 of Quang 2014. Optional: Bishop Textbook, Sec. 2.3.9, 9.2-3 |
Project 3 P3-Data HW10 |
| 25/02/1398 | Lecture 11- Undirected graphical models; variable elimination (slides) | Required: Koller Textbook, Sec. 4.1-2, 9.2, 9.3.1 | HW11 |
| 1/03/1398 | Lecture 12- Variable elimination; Gibbs sampling for motif finding (slides); Gaussian mixture model (notes) | Koller Textbook, Sec. 9.3.2, 9.4.2; Compeau Textbook, Ch. 2 | HW12 |
| 08/03/1398 | Lecture 13- Clique trees; MAP inference (slides) | Koller Textbook, Sec. 10.1, 10.2.1-2, 10.4, 13.1.2, 13.2 | |
| 08/03/1398 | Lecture 14- Bayesian networks for temporal data (slides) (notes) | Koller Textbook, 6.1-2; Compeau Textbook, Ch. 10 | HW13/14 Data R code |
| 13/03/1398 | Lecture 15- Particle-based approximate inference (slides) | Koller Textbook, Sec. 12.1-3 | HW15 |