14. Expert Systems

Expert Systems

What is an Expert System?

An expert system is a computer program that emulates the decision-making ability of a human expert in a specific domain. It was one of the earliest successful applications of AI, booming in the 1980s.

Architecture

• Knowledge Base — Contains domain-specific facts and rules (typically IF-THEN rules)
• Inference Engine — Applies rules to known facts to derive new conclusions
• User Interface — Allows users to query the system and receive explanations
• Explanation Module — Explains how the system reached a conclusion ("Why?" and "How?" )
• Knowledge Acquisition Module — Helps experts add new knowledge to the system

How It Works

IF the patient has a fever
AND the patient has a cough
AND the patient has been exposed to COVID-19
THEN the patient likely has COVID-19 (confidence: 90%)

IF the patient likely has COVID-19
AND the patient has difficulty breathing
THEN recommend immediate hospitalization

Inference: Forward Chaining
- Start with known facts (symptoms)
- Apply rules to infer intermediate conclusions
- Continue until goal is reached (diagnosis)

Famous Expert Systems

• MYCIN — Medical diagnosis of bacterial infections (1970s)
• XCON/R1 — Computer system configuration for DEC (1980s, saved $40M/year)
• DENDRAL — Chemical structure analysis from mass spectrometry data
• PROSPECTOR — Mineral exploration, discovered a molybdenum deposit worth $100M

Advantages

• Consistent, unbiased decision-making
• Preserves expert knowledge even after expert leaves
• Available 24/7, can be replicated and distributed
• Can explain reasoning steps

Limitations

• Limited to narrow domains
• Difficult to maintain and update knowledge base
• Cannot handle novel situations outside its rules
• Knowledge acquisition bottleneck — extracting expert knowledge is hard
• No learning capability — rules must be manually updated