[an error occurred while processing this directive]
COMPSCI 276 Fall 2007, Network-Based Reasoning - Bayesian Networks
[ main | homework | handouts | readings | software | links | projects| books ]

Projects - pdf

Course Reference

  • Days: Monday/Wednesday
  • Time: 2:00 p.m. - 3:20 p.m.
  • Room: DBH 1423
  • Instructor: Rina Dechter
  • Office Hours: Wednesday 3:30-4:30 p.m. Location: Bren Hall 4232

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 augmenting probablistic networks with constraints and on exact and approximate probabilistic reasoning algorithms. Additional topics include: causal networks, first-order probablistic languages and dynamic Bayesian networks.


  • 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.

Tentative Syllabus

Date Topic Readings
Week 1 10/01
  • Welcome; Introduction to Bayesian networks.
(a) Pearl chapter 1-2,
(b) Russell-Norvig chapter 13
Homework 1
Class Slides 1
  • Probabilistic networks representation: Independence properties.

Week 2 10/08
  • Markov networks: Undirected graphical models of independence. Knowledge engineering.
(a) Pearl chapter 3
(b) Optional: Koller Chapter 3 (up to 3.4.2)
Homework 2
  • Bayesian networks: Directed graphical models of independence.
(a) Koller chapter 5

Week 3 10/15
  • Reasoning with Probabilistic networks:
  • Exact reasoning by inference: Variable elimination.

  • Tree-decompositions: bucket trees, join-trees and polytrees. Cluster tree elimination and propagation algorithms.
(a) Bucket Elimination: A Unifying Framework for Probabilistic Inference
(b) Class Notes (1-2)

Week 4 10/22
  • Reasoning with probabilistic networks (continued).
(a) Koller chapter 8
Homework 3
Due 10/29
Class Slides 2
  • Reasoning with Probabilistic networks (continued).

Week 5 10/29
  • Inference augmented with simple search: The loop-cutset and w-cutset schemes,
  • The MPE and MAP queries.

Homework 4
Due 11/07
Class Slides 3 (continued)
  • Approximate reasoning by bounded inference: mini-bucket, mini-clustering, belief propagation schemes.

Week 6 11/5
  • Approximate reasoning by sampling: MCMC methods (Gibbs sampling), importance sampling.
  • Custet conditioning sampling.
(a) Pearl, Chapter 4
(b) Cutset sampling in Bayesian networks
(c) Koller, Chapter 10
Homework 5
Due 11/14
Class Slides 4 (a)
  • Sampling (continued)

Class Slides 4 (b)
Week 7 11/12
  • No class. Veteran's day holiday

  • Representation: Local structures. Causal independance, context-specific and determinism

Week 8 11/19
  • Reasoning Hybrids trading Time and Space: AND/OR Branch and Bound and Best-first with mini-bucket heristics, AND/OR w-cutset, VEC.
(a) Class Notes (3)
(b) AND/OR Search Spaces for Graphical Models
(b)AND/OR search space for graphical models
Homework 6
Due 11/28
Class Slides 5
  • Extended representations: Mixed probabilisic and deterministic networks. Hybrid discrete and continuous networks. Dynamic Bayesian networks. First-order probabilistic languages.

Class Slides 6
Week 9 11/26
  • Causal networks

Class Slides 7
Class Slides 8
Class Slides 9
  • Causal networks (continued)

Week 10 12/3
  • Project presentations.

  • Project presentations.


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

Grading Policy:

Homeworks and projects (70%), midterm (30%)