Ira J Kalet
A readings class in knowledge representation, as described in the primary artificial intelligence and biomedical informatics literature. Topics may include: frame-based systems, description logics, theorem proving, complexity vs. tractability, ontologies, rule-based systems, and biomedical domain applications. Prerequisite: any artificial intelligence course (e.g., CSE 415 or better), or permission of instructor. Offered: A.
The class is more didactic than the catalog description suggests, with textbooks as well as supplemental materials. Students can expect to develop skill at formal proofs, and the use of programming techniques to create, maintain and use examples of systems and methods mentioned in the topics list. Biomedical examples will be used throughout, so that students will be able to apply knowledge representation ideas and formalisms specifically to biomedical problems. Students will be able to identify and analyse challenges in biomedical knowledge representation pointing to current research.
Student learning goals
Distinguish among various kinds of data representations and encodings used in biomedicine and give examples of the use of each,
Translate biological, medical or public health factual assertions or statements of knowledge into formal logic using traditional logic notation and in a form that can be used directly by a computer program or system,
Describe basic properties of frame systems and write frame representations for examples of biomedical knowledge,
Perform subsumption and classification of statements written in a description logic,
Apply basic Bayesian probability formulas to answer questions about probabilistic biomedical problems,
Analyze biomedical knowledge representation and reasoning problems and give reasons for using or not using particular programming languages, software environments or methodologies.
General method of instruction
The class will meet twice a week for an hour and twenty minutes. Since the class is small, there will be a mix of didactic presentation and in-class discussion. Questions and comments will always be welcome. Additional optional sessions may be scheduled if needed.
The catalog prerequisite is out of date. A course in artificial intelligence is NOT required. The basic requirement is some experience with computer programming, scripting languages, text editors, and the general environment of programming. However, it is not important which programming language you know, as long as you have experienced at least one. In addition, the course requires basic computing skills and mathematical background.
Class assignments and grading
Weekly assignments will be taken from the required textbooks, and additional exercises will be assigned by the instructor. Some will involve formal and informal proofs, some will involve programming, and some will involve the use of software packages that will be available in the Informatics Lab of the Biomedical and Health Informatics graduate program. Some parts of the assignments may be in essay format. Each student will also be expected to do a project involving programming or other technical work, submit a written report and do a short presentation of their project at the end of the quarter.
The assignments will be graded and will make up the majority of the grade. The project will be graded based on both written and oral presentation. Class participation will be factored into the final grade.