CIS 603. Artificial Intelligence

Knowledge Management

1. The Problem

2. Proposed Solutions

structure: semantic network, frame, script, description logic, ontology

psychological models of categorization: classical theory, prototype theory, exemplar theory, theory theory

"strong methods" vs. "weak methods": content vs. mechanism

expert system (knowledge-based system), rule-based vs. case-based

knowledge base: Cyc, Open Mind, WordNet, Mizar

knowledge acquisition: knowledge engineering, semantic web, data mining, natural language interface

3. Issues

efficiency

4. Readings

5. Ideas

reasonable:

• domain knowledge matters
• knowledge based requires structure

• knowledge is given or built-in
• knowledge structure is static

• memory structure described in The Logic of Categorization
• multiple knowledge sources