Organized Outline

1. Introduction

• Expert Systems (ES) can perform at a level equivalent to that of a human expert.
• ES is highly domain-specific, applicable in electronics, law, manufacturing, medicine, and chemistry.
• Adequate response time, ability to explain reasoning, and handling uncertainties are key features.

2. Definition

• ES is a set of computer programs that can advise, consult, diagnose, explain, forecast, interpret, justify, learn, plan, etc.
• An “expert system” is defined as a computerized clone of a human expert.

3. Expert and ES Shell

• Expert
  • Possesses considerable knowledge in a specific field.
  • Domain-specific.
• ES Shell
  • A special-purpose tool designed for specific applications.
  • Isolates knowledge bases from the reasoning engine, improving software portability.

4. Shell Concept for Building Expert Systems

• Components: KB, Consultation Manager, KB Editors & Debugger, Explanation Program, KBMF Inference Engine Shell.
• Separation of knowledge base and inference mechanism.

5. Characteristics: Comparison

• Conventional Systems
  • Information and processing combined.
  • No mistakes, but programmers can make errors.
  • Execution is step-by-step.
• Expert Systems
  • Knowledge base separated from processing.
  • Can operate with incomplete information.
  • Changes in rules are easy to accomplish.

6. Early ES (70s – mid-80s)

• Examples: MYCIN, DENTRAL, PROSPECTOR, XCON, REVEAL, CENTAUR, HEARSAY I.
• Characteristics common to early ES.

7. Generic Categories of ES

• Interpretation, Prediction, Diagnosis, Design, Planning, Repair, and other categories.

8. Benefits

• Expertise in a field made available to more people.
• Preserves top experts’ knowledge.
• Facts stored systematically.
• Allows handling more complex problems.

9. Limitations

• Systems may be too superficial.
• Rapid degradation of performance.
• Crude interfaces.
• Inability to adapt to multiple types of reasoning.

10. Basic Structure of a Production System

• Components: Production Rules, Knowledge Base (KB), Database (DB), User, Inference Engine, Explanation Facility, User Interface.

11. ES Structure

• Components: Consultation Environment, Development Environment, User Interface, Inference Engine, Explanation Facility, Working Memory, Facts of the Case, Recommendation.

12. Key Components of ES

• Inference Engine, Explanation Module, User Interface, Working Memory, Knowledge Base, Knowledge Acquisition Facility.

13. Explanation Module/Facility/Justifier

• Importance of Explanation.
• Approaches: Canned Text and Paraphrase.

14. Inference Engine

• Forward Chaining and Backward Chaining.
• Deduction of conclusions and determination of causes.

15. Summary

• Recap of key points.

16. Rule-based Validation

• Identifying inconsistencies: Redundant Rules, Conflicting Rules, Subsumed Rules, Unnecessary Premise (IF) Clauses, Circular Rules.