Person/Douglas Lenat
About
- American computer scientist and artificial intelligence researcher who founded and led Cycorp.
- Known for the AM, Eurisko, and Cyc projects, and for pioneering work in knowledge representation and ontological engineering.
- Born September 13, 1950; died August 31, 2023.
Knowledge principles
- In On the Thresholds of Knowledge, Douglas Lenat and Edward Feigenbaum argued that intelligent systems need substantial, explicit knowledge about the world.
- Knowledge Principle: competent performance depends primarily on the concepts, facts, representations, methods, models, metaphors, and heuristics a system can bring to bear—not on search and reasoning alone.
- Explicit Knowledge Principle: knowledge may be compiled into opaque code for efficiency, but a declarative form should remain available for inspection and meta-reasoning.
- Breadth Hypothesis: intelligent behavior in unexpected situations requires fallback to increasingly general knowledge and analogy to specific knowledge from distant domains.
- Empirical Inquiry Hypothesis: investigate AI by building programs and testing them on large, reality-bearing problems; premature mathematization and toy problems can hide important details.