Introduction
(9/8)
Motivation and background
Qualitative and quantitative approaches to reasoning under uncertainty
Probabilistic reasoning
Probabilistic representation of uncertainty
(9/13) Reading: [Russell and Norvig, Chapter 13]
Probability distributions
Prior and conditional probability
Inference using joint distributions
Conditional independence and Bayes' rule
The semantic of Bayesian networks
Conditional independence relations
Efficient representation of conditional distributions
(Notes)
Dependency models and maps
(9/15) Reading: [Castillo et al, 5, 6] [Pearl, 1988, 3.1]
Qualitative reasoning about independence relationships
Dependency models and dependency maps
(Notes)
Bayesian networks
(9/20) Reading: [Korb and Nicholson, 2] [Russell and Norvig, 14.1-14.3]
The semantic of Bayesian networks
Efficient representation of conditional distributions
Reasoning with Bayesian networks
(Notes)
Exact inference in BNs
(9/22, 9/27) Reading: [Castillo et al, 8] [Russell and Norvig, 14.4]
The complexity of exact inference
Inference by enumeration
Pearl's message passing algorithm
The variable elimination algorithm
Clustering methods
Junction trees
(Notes)
(Notes)
Approximate inference in BNs
(9/29) Reading: [Russell and Norvig, 14.5] [Korb and Nicholson, 3.6]
Approximate inference with stochastic simulation
Direct sampling methods, rejection sampling and likelihood weighting
Markov Chain Monte Carlo (MCMC)
(Notes)
Causal inference
(10/4) Reading: [Pearl, 2002] [Halpern and Pearl, 2001, Part I]
Reasoning about cause and effect
Causes and explanations
(View Judea Pearl's talk about Causal Inference)
Decision networks
(10/6, 10/13) Reading: [Russell and Norvig, 2004, 16] [Qi and Poole, 1995]
The basis of utility theory
Decision trees and influence diagrams
The value of perfect and imperfect information
Evaluating influence diagrams
(Aron Culotta)
(Notes)
(Notes)
Probabilistic reasoning over time
(10/18) Reading: [Russell and Norvig, 2004, 15.1-15.5]
Inference in temporal models
(Jamie Rothfeder)
Dynamic Bayesian networks
Inference algorithms for DBNs
(Notes)
Applications of BNs and knowledge engineering
(10/25, 10/27) Reading: [Korb and Nicholson, 9-10]
Knowledge engineering with BNs
(Kedar Bellare)
Evaluation and validation methods
(Nadia Ghamrawi)
(Notes)
(Notes)
Sequential decision making
(11/1, 11/8) Reading: [Russell and Norvig, 2002, 17.1-17.3] [Hansen and
Zilberstein, 2001] (Optional: [Feng and Zilberstein, 2004])
Markov decision processes
Value iteration and policy iteration
Solving MDPs using heuristic search
Solving Partially observable MDPs
(Sven Seuken)
(Notes)
(Notes)
Structured representations
(11/10) Reading: [Hoey et al, 2000] [Feng and Hansen, 2002]
Exploiting domain structure to improve the efficiency of inference
Stochastic Planning Using Decision Diagrams
(Siddharth Srivastava)
Symbolic Heuristic Search for Factored Markov Decision Processes
(Guest lecture: Zhengzhu Feng)
(Notes)
Multiple agents acting under uncertainty
(11/15, 11/17) Reading: [Bernstein et al, 2002]
[Russell and Norvig, 2004, 17.6-17.7]
[Koller and Milch, 2003]
Modeling collaborative multi-agent systems as decentralized MDPs
(Guest lecture: Daniel Bernstein)
Game theoretical approaches
(Chris Amato)
Graphical models for games
(Daniel Bernstein)
(Notes for 11/15)
(Notes for 11/17)
Variational methods
(11/29) Reading:
[Jordan et al, 1999]
Introduction to Variational Methods for Graphical Models
(Guest lecture: Ramesh Nallapati)
(Notes)
Bounded rationality
(12/1) Reading:
[Simon, 1982] [Russell, 1997] [Hansen and Zilberstein, 2001 (Monitoring)]
Acting under uncertainty with limited computational resources
The value of computation
Monitoring and control of anytime algorithms
Models of bounded rationality
Handling uncertainty in computation
(Ron Bekkerman)
(Notes)
Non-Bayesian quantitative approaches
(12/6, 12/8) Reading: [Russell and Norvig, 2004, 14.7] [Giarratano and Riley, 5.4-5.5]
Rules-based methods
Fuzzy sets, fuzzy logic and possibility theory
(Paul Dickson)
The Dempster-Shafer theory of evidence
(Kimberly Ferguson)
(Notes for 12/6)
(Notes for 12/8)
Applications of probabilistic reasoning
(12/13) Reading: [Thurn, 2000]
Probabilistic Approaches to Mapping and Localization in Mobile Robots
(Steve Hart)
Course summary
(Notes)