Overview
This lecture covers the emergence of fifth-generation computers, focusing on knowledge-based (expert) systems, their development, applications, the underlying artificial intelligence (AI) technologies, and the global competition in this field.
Fifth Generation and Expert Systems
- Fifth-generation computing refers to advanced knowledge-based or AI-driven systems, sometimes called Knowledge Information Processing Systems (KIPS).
- The Japanese government has initiated a major national program to lead in knowledge-based computing, aiming to leap ahead in global markets.
- Knowledge-based systems are designed to emulate or surpass human expert reasoning using encoded field knowledge.
Foundations of Knowledge-Based Systems
- Expert systems process symbolic knowledge and use inference, not just numerical calculation.
- Two key knowledge types: factual (documented, agreed-upon) and heuristic (experiential, intuitive).
- Knowledge engineering is the process of encoding both factual and heuristic knowledge into a system.
- The main power of expert systems is the depth and breadth of knowledge encoded, not just logical inference methods.
- The shift to a knowledge-based paradigm is the most significant development in AI's recent decades.
Applications of Expert Systems
- Early expert systems included chemical analysis (DENDRAL) and medical diagnosis (MYCIN, Pittsburgh's Internal Medicine program).
- Applications have expanded into equipment diagnosis, computer configuration, integrated circuit design, and more.
- Example: The Drilling Advisor system helps oil companies diagnose and solve drilling problems worldwide using captured expert knowledge.
How Expert Systems Work
- Many expert systems are rule-based: knowledge is expressed as "if-then" rules.
- They employ backward chaining, working from a goal backward to known facts.
- Expert systems can explain their reasoning to users, promoting transparency and trust by tracing rule chains to justify advice.
- These systems are intended as decision-support tools for professionals.
U.S. and Japanese Competition in AI
- Japan is heavily funding fifth-generation projects to gain an early lead, especially in areas where IBM and other U.S. firms are less active.
- The U.S. has historically led in expert systems but faces the need to accelerate its efforts to stay ahead.
Key Terms & Definitions
- Fifth Generation Computer — Advanced computers centered on AI and knowledge-based systems.
- Expert System — A software program emulating human expert decision-making in a specific field.
- Knowledge Engineering — The practice of encoding expert knowledge into a computer system.
- Heuristic Knowledge — Experiential, intuitive know-how used by human experts.
- Rule-Based System — An expert system using "if-then" logic to reach conclusions.
- Backward Chaining — An inference strategy that works from a desired goal back to available facts.
Action Items / Next Steps
- Review examples of expert systems in medicine, engineering, and business.
- Read about the Japanese fifth-generation project for further context.
- Explore how rule-based systems use backward chaining in problem-solving scenarios.