Course Reference

  • Days: Tuesday/Thursday
  • Time: 11:00 p.m. - 12:20 p.m.
  • Room: DBH 1423
  • Instructor: Rina Dechter
  • Office hours: TBA

Course Description

One of the main challenges in building intelligent systems is the ability to reason under uncertainty, and one of the most successful approaches for dealing with this challenge is based on the framework of Bayesian networks, also called graphical models. Intelligent systems based on Bayesian networks are being used in a variety of real-world applications including diagnosis, sensor fusion, on-line help systems, credit assessment, bioinformatics and data mining.

The objective of this class is to provide an in-depth exposition of knowledge representation and reasoning under uncertainty using the framework of Bayesian networks.  Both theoretical underpinnings and practical considerations will be covered, with a special emphasis on dependency and independency models, on construction Bayesian graphical models and on exact and approximate probabilistic reasoning algorithms. Additional topics include: causal networks, learning Bayesian network parameters from data and dynamic Bayesian networks.

Prerequisites

  • Familiarity with basic concepts of probability theory.
  • Knowledge of basic computer science, algorithms and programming principles.
  • Previous exposure to AI is desirable but not essential.

Course material

The course will be based mostly on three sources:

Additional sources:

A longer list including secondary references.

Some links to software and tools.

Tentative Syllabus

Week     Date Topic Readings           Files
Week 1 3/29
  • Introduction to Bayesian networks.
(a) Pearl 1-2
(b) Darwiche 1-3
(c) Russell-Norvig 13
(d) Darwiche. Bayesian Networks
Homework 1
Class Slides 1
  3/31
  • Probabilistic networks representation: Independence properties.


Week 2 4/5
  • Markov networks: Undirected graphical models of independence.

Homework 2
Class slides 2
Lecture 2
  4/7
  • Bayesian networks: Directed graphical models of independence.
Pearl Ch.3
Class slides 3
Lecture 3
Week 3 4/12
  • Bayesian networks: Directed graphical models (continued)

Class slides 4
Lecture 4
  4/14
  • Building Bayesian networks

Class slides 5
Lecture 5
Week 4 4/19
  • Exact inference: Variable elimination
Reasoning with graphical models. Chapters 1-5

Homework 3

4/21
  • Optimization queries: MPE and MAP


Week 5 4/26
  • Tree-decompositions: bucket trees, join-trees and polytrees. Cluster tree elimination and propagation algorithms.

Lecture 6

4/28
  • Tree-decompositions: bucket trees, join-trees and polytrees. Cluster tree elimination and propagation algorithms, continued.

Lecture 7
Week 6 5/3
  • Search and Inference: The loop-cutset and w-cutset schemes.
Reasoning with graphical models: Chapter 6

Class slides 6
Lecture 8 Homework 4

5/5
  • AND/OR search spaces

Class slides 7
Lecture 9
Week 7 5/10
  • Approximate reasoning by sampling: MCMC methods (Gibbs sampling), importance sampling.

Lecture 10

5/12
  • Custet conditioning sampling.


Week 8 5/17
  • Sampling continued
Reasoning with Graphical Models: Chapter 7
Lecture 11
Homework 5
5/19
  • Approximate inference: mini-buckets, IJGP

Lecture 12, bounding inference
Lecture 12, sampling
Class slides 8
Week 9 5/24
  • Mini-buckets continued
Reasoning with Graphical Models: Chapter 8
Homework 6
Lecture 13
Class slides 9

5/26
  • Mini-buckets continued

Lecture 14

Week 10 5/31
  • Project presentations.



6/2
  • Project presentations.



Assignments:

There will be homework assignments and students will also be engaged in projects.

Grading Policy:

Homework and exam (75%), class project (25%)