REASONING IN INTELLIGENT SYSTEMS (fall 2000)

The course is for 3rd year undergraduates and graduate students of CS. We will mostly study the problem of reasoning in presence of uncertainty. Different approaches to this problem are considered: Certainty Factor calculation, Probabilistic approach, Dempster-Shafer theory of evidence, Fuzzy Logic etc. In this connection different aspects of Learning Theory are discussed. In the last part the new development connected with these problems: Knowledge Discovery in Databases is considered.

Course plan:

• Historical background: Artificial Intelligence, Problems of Reasoning and Applications
• Precise Reasoning. Logic and Theorem Proving. Rules of Inference, Resolution Proofs.
• Reasoning in Rule-Based, Frame-Based and Semantic Net-Based Systems. (Review)
• Problems with application of logic based approach to Common Sence Reasoning. Reasoning with Uncertainty. Certainty Factor.
• Probabilistic Reasoning, Pooling of Evidences. Recursive Bayessian Updating.
• Bayessian Probabilistic Network (BPN). Representation of BPN in the form of DAG. Connection between structure of DAG and properties of BPN.
• Abductive Reasoning and Explaining Away in BPN.
• Learning probabilities in BPN.
• BPN and Markov Random Fiel (MRF). Comparison and Applications.
• Dempster-Shafer Theory of Evidence.
• Probabilistic Reasoning and Decision Making. Risk Function and Utility Function based approaches.
• Fuzzy Logic and Fuzzy Logic based controllers. Fuzzy Logic and Probabilistic Reasoning.
• Inference in Fuzzy Logic based systems.
• Learning in Intelligent Systems. Inductive Reasoning (IR).
• IR with Tree-like structures. Cluster Analysis and IR.
• IR and Artificial Neural Nets.
• Knowledge Discovery in Databases. Requirements and Methods.
• Review of some optimization methods (Sumulating Annealing, Genetic Algorithms, Evolutionary Programming) and their application to IR.

Assignements: 6 homeworks with exercises. At the end of the course the students should pass exams in the form of a special homework (during three days). The final mark is mostly based on the result of this final homework. The results of mentioned 6 homeworks also have an influence on the final mark.

Requirements: Standard courses on Mathematics (Probability and Statistics, Linear Algebra, Infi).

Literature:

P.H. Winston, Artificial Intelligence, Addison-Wesley, 1984.

J. Pearl, Probabilistic Reasoning in Intelligent Systems, Morgan Kaufmann, 1988.

A.Kandel, Fuzzy Expert Systems, CRC Press, 1991.

Editors: U.M. Fayyad et others, Advances in Knowledge Discovery and Data Mining, AAAI Press/The MIT Press, 1996.