Expert system pros and cons

An expert system is a computer program that stores the knowledge of human experts in a specific domain as facts and rules and uses an inference engine to give expert-level advice, diagnosis or decisions (e.g. MYCIN for medical diagnosis).

Main components: knowledge base, inference engine, working memory, user interface, explanation facility and knowledge acquisition facility.

Advantages

1. Availability: works 24 hours a day, anywhere it is installed.
2. Preserves expertise: keeps knowledge when experts retire or leave, and makes scarce expertise widely available.
3. Consistency: gives the same answer for the same case; no fatigue, emotion or bias.
4. Speed: gives answers faster than a human consultation.
5. Lower cost: reduces the cost of expert consultation after development.
6. Explanation: can explain its reasoning (How? Why?), which helps trust and training.
7. Handles complex rules: can combine large numbers of rules and facts without forgetting any.
8. Works in hazardous places where humans cannot go.
9. Training aid for novices.

Disadvantages

1. High development cost and time, especially knowledge acquisition (the bottleneck).
2. No common sense or creativity: works only in a narrow domain and fails outside it.
3. Cannot learn automatically from experience (most systems need manual updates).
4. Brittle: gives wrong or no answers for cases not covered by the rules.
5. Hard to extract knowledge: experts often cannot express their intuition as rules.
6. Maintenance is hard as rule bases grow; rules may conflict.
7. Lack of human touch and responsibility: legal and ethical issues if the advice is wrong.
8. Depends on input quality: needs reliable inputs; errors in the data give errors in the output.