home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.ai
- Path: sparky!uunet!europa.asd.contel.com!darwin.sura.net!haven.umd.edu!news.umbc.edu!algol.cs.umbc.edu!finin
- From: finin@algol.cs.umbc.edu (Timothy Finin)
- Subject: Models Of Diagnostic Inferences
- Message-ID: <1992Sep2.203337.288@umbc3.umbc.edu>
- Sender: newspost@umbc3.umbc.edu (News posting account)
- Organization: Computer Science, University of Maryland, Baltimore County
-
- GRADUATE AI COURSE ANNOUNCEMENT
- Distribution: md,dc
- Date: Wed, 2 Sep 1992 20:33:37 GMT
- Lines: 59
-
- Computer Science Department
- University of Maryland Baltimore County
- Baltimore MD
-
- CMSC 691: MODELS OF DIAGNOSTIC INFERENCES
-
- Fall, 1992, TuTh 4:00 - 5:15pm
-
-
- Instructor Yun Peng
- Phone: (301)455-3816
- Office: TF II, room 105
- Email: ypeng@algol.cs.umbc.edu
-
- Text
- Y. Peng & J.A. Reggia, "Abductive Inference Models for
- Diagnostic Problem-Solving", Springer-Verlag, 1990, and
- supplementary journal papers.
-
- Course Description
- This course offers an introduction and in depth discus-
- sion of important issues in the area of abduction, a
- prominent but less studied type of inference, and diag-
- nostic reasoning, a typical representative of abductive
- inference. Topics include the nature and the applica-
- tions of abduction; knowledge representation and infer-
- ence paradigms for diagnostic reasoning; formal and
- heuristic methods for uncertainty in causal reasoning,
- etc. Several state-of-the-art AI and neural netwrok
- models for diagnostic problem-solving will also be dis-
- cussed.
-
- Major Topics To Be Covered
- 1. Abduction vs deduction and induction, its unique
- characteristics, and its applications.
- 2. Different representations for causal knowledge and
- the characteristics of causal inference.
- 3. Uncertainty in causal reasoning and representative
- approaches to this problem.
- 4. Acquisition of causal knowledge (learning).
-
- Theoretical Models To Be Discussed
- 1. Shortliffe et al: MYCIN
- 2. Peng & Reggia: Parsimonious Covering Theory and Pro-
- babilistic Causal Model
- 3. Reiter: Theory of Diagnosis From the First Principle
- 4. De Kleer: Truth Maintenance System and the General
- Diagnostic Engine
- 5. Pearl: Bayesian Network
- 6. Dempster-Shafer theory
- 7. Neural networks: Backpropagation networks; Hebbian
- learning method
-
- Work Load
- One exam (50%), one paper/project (50%).
-
- Background
- Necessary: Basics of AI, elementary set theory and probability theory.
- Desirable: Basic knowledge of neural networks.
-