Uncertainty in AI Seminar

Description of the course

The course will consist of an introduction to the course topics, followed by student talks on selected papers. The outline of the first part of the course is as follows.

1. Introduction
   • Why uncertainty?
   • Survey of systems for uncertain reasoning
2. Evidential (probabilistic) reasoning
   • Probabilistic updating - Markov and Bayes networks
   • Decision theory and influence diagrams
   • Ordinal probabilities ("kappa calculus")
   • Probabilistic logic
   • Belief functions (Dempster-Shafer Probabilities)
3. Other formalisms
   • Non-monotonic logics
   • Fuzzy logic
   • Connectionism and coarse coding
4. Applications
   • Commonsense reasoning
   • Natural language understanding
   • Vision
   • Planning and robot control

Reading List

1. "Probabilistic Reasoning in Intelligent Systems", Judea Pearl, Morgan Kaufmann
2. "Readings in Uncertain Reasoning", Shafer and Pearl, Morgan Kaufmann
3. Articles from the uncertainty in AI community, and related papers.
   • Proceedings of the Conference on Uncertainty in AI, 1990-1996
   • Selected papers from International Jornal of Approximate Reasoning
   • Selected papers from Artificial Intelligence Journal

Requirements and Prerequisites

This course is predominantly for graduate students, but senior year (3rd year or above) CS students may also attend. Atttendance of all classes, or an equivalent replacement, is mandatory. One 2-hour talk on a set of articles from the reserach literature will be prepared and presented by each student.

Administrative Issues

• Instructor: Dr. S. E. Shimony shimony@cs.bgu.ac.il
• Credits: 1 credit for both graduate and undergraduate students.

Back to BGU Math & CS HomePage