Lecture on Fifth-Generation Computers and Expert Systems
Overview
- Presenter: Stuart Shafie
- Guest Expert: Edward Feigenbaum
- Topic: Fifth-generation computers, AI, expert systems
- Alternate Names: Knowledge Information Processing Systems (KIPS), Knowledge-based systems
Key Concepts
Fifth-Generation Computing
- Represents a quantum leap in computing technology
- Japanese Initiative: Ten-year national program focusing on KIPS
- Central to the fifth generation: Knowledge-based systems (Expert Systems)
Expert Systems
- Definition: Systems infused with knowledge to provide intelligent solutions
- Components:
- Documented facts
- Heuristic or experiential knowledge
- Method: Knowledge engineering
Knowledge Engineering
- Process of embedding human-like experiential knowledge into systems
- Systems can manipulate facts and make decisions
- Goal-directed backward chaining used for inference
Historical Context
Evolution of Computing
- Evolution marked by different generations, with the fifth generation focusing on AI and knowledge processing
- Influences: Alan Turing, Charles Babbage, Ada Lovelace
- Artificial Intelligence (AI): Focused on processing symbolic knowledge using logical inference
Milestones in AI
- Early AI research focused on generality and symbolic manipulation
- Shift to Knowledge-Based AI: Emphasis on embedding substantial domain-specific knowledge
- George Polya (Stanford Mathematician): Popularized heuristic reasoning
Practical Applications
Examples of Expert Systems
- Chemical Analysis: DENDRAL program at Stanford (1965)
- Medical Diagnosis: MYCIN program for infectious disease diagnosis & therapy
- Industrial Applications:
- Configuring customer orders
- Designing integrated circuits
- Diagnosing computer equipment failures
Industry Transition
- Transition from universities to industry (similar to late 1940s and early 1950s computing revolution)
- Applications across various sectors, especially engineering and healthcare
Specific Case Study: Drilling Advisor
- Problem: Addressing drilling issues in oil wells remotely
- Developer: ELF Aquitaine (French National Oil Company)
- Prototype Demo:
- System asks for specific data inputs and advises on drilling issues
- Rule-based system: Uses rules with if-then logic
- Explanation system allows transparency and understanding of the reasoning process
International Competition: Japan's Fifth-Generation Project
- Goal: Establish industry leadership in AI and knowledge systems
- Funding: $450 million government funding plus matching industry funding
- Strategy: Capture market share in AI where leading companies like IBM are less active
- Call to Action: US needs to accelerate efforts to maintain leadership
Industry News and Developments
- Semiconductor Industry: Effort to regain dominance, development of 4-megabyte chip by 1988
- Fortune Computers: Issues leading to resignation of CEO Gary Friedman
- MCI Mail: New electronic mail service by MCI
- Portable Computers on Airlines: Restrictions on use due to potential radio interference
- Nolan Bushnell (Founder of Atari): New ventures in video games and robotics
- Legal Rulings: Apple vs. Franklin Computers – software on silicon chips deemed copyrightable
- Best-Selling Business Software: Lotus 1-2-3
- Advertising in Software: Companies planning to place ads on software discs and computer networks
Conclusion
The lecture provides a comprehensive overview of the advancements and implications of fifth-generation computing, focusing on the development and application of expert systems, and highlighting the need for strategic advancement to stay competitive globally.