Artificial Intelligence & Expert Systems

Artificial Intelligence (AI) represents one of the most transformative frontiers in modern computer science. One of the earliest and most commercially successful branches of applied AI is the Expert System—software engineered to replicate the problem-solving and diagnostic capabilities of a human domain expert.

Learning Objectives

• Define Artificial Intelligence (AI) and understand its core domains.
• Describe the architecture and components of an Expert System: User Interface, Knowledge Base, Rule Base, Inference Engine, and Explanation Facility.
• Trace how an expert system uses forward chaining to deduce conclusions from input data.
• Evaluate the advantages and limitations of expert systems compared to human specialists.

What Is Artificial Intelligence?

Artificial Intelligence (AI) is the branch of computer science dedicated to developing computer systems capable of performing tasks that traditionally require human intelligence. These tasks include:

• Visual and auditory perception (facial recognition, speech-to-text).
• Natural language comprehension and translation.
• Complex decision-making and pattern recognition.
• Autonomous navigation and problem-solving.

What Is an Expert System?

An Expert System is an AI program designed to emulate the decision-making and reasoning ability of a human expert in a specialized domain:

• Medical Diagnosis: Evaluating complex patient symptoms to recommend treatment options (e.g. MYCIN).
• Mineral & Petroleum Exploration: Analyzing geological soil samples and seismic data to locate underground oil or mineral deposits.
• Automotive Fault Diagnostics: Troubleshooting complex electrical and mechanical issues in aircraft or vehicles.
• Financial & Tax Advisory: Calculating tax liability and evaluating business loan eligibility.

Components of an Expert System

An expert system consists of interconnected functional subsystems:

+-------------------------------------------------------------+
|                         HUMAN USER                          |
|                     (Doctor, Engineer)                      |
+-------------------------------------------------------------+
                              ^
                              | (Queries & Explanations)
                              v
+-------------------------------------------------------------+
|                       USER INTERFACE                        |
|             Presents questions, displays diagnosis          |
+-------------------------------------------------------------+
                              |
                              v
+-------------------------------------------------------------+
|                      INFERENCE ENGINE                       |
|   Applies Rules to Facts using Forward/Backward Chaining    |
+-------------------------------------------------------------+
       |                                             |
       v                                             v
+---------------------------+         +---------------------------+
|      KNOWLEDGE BASE       |         |         RULE BASE         |
| Factual Database of Data  |         | IF... THEN Logical Rules  |
| gathered from specialists |         | developed by experts      |
+---------------------------+         +---------------------------+
       ^
       |
+---------------------------+
|   EXPLANATION FACILITY    |
| Displays reasoning trail  |
+---------------------------+

1. User Interface

• Facilitates communication between the non-expert user and the system.
• Poses targeted questions regarding symptoms, measurements, or conditions.
• Displays the final conclusion, diagnosis, or recommended action along with a confidence rating (e.g. "85% probability of Condition X").

2. Knowledge Base

• A vast, structured repository of factual information, statistics, reference tables, and domain-specific knowledge compiled from recognized human experts through knowledge engineering.

3. Rule Base

• A comprehensive collection of logical IF... THEN inference rules capturing the heuristic reasoning and instincts of human experts.
• Example:

IF (Engine_Turns_Over = TRUE) AND (Spark_Plug_Fires = FALSE) THEN (Fault = Ignition_Coil)

4. Inference Engine

• The central reasoning engine (the "brain") of the expert system.
• Acts as an algorithm that systematically evaluates user inputs against the Knowledge Base and fires applicable rules in the Rule Base to infer new facts and deduce conclusions.

5. Explanation Facility / System

• Explains the reasoning behind the system's deductions.
• Answers two critical user questions:
• "Why?": Explains why the system is asking a specific question.
• "How?": Demonstrates the step-by-step logical chain of rules and facts used to arrive at the final diagnosis.
• Vital in high-stakes fields (like medicine and law) where clinicians cannot accept recommendations without understanding the justification.

Advantages & Limitations of Expert Systems

Written practice and self-review

A medical clinic deploys an Expert System to support junior clinicians in diagnosing complex infectious illnesses.

1. Describe how the Knowledge Base, Rule Base, and Inference Engine work together when a clinician inputs a patient's symptoms.
2. Explain the purpose of the Explanation Facility / System and discuss why it is clinically vital.

Write an explanation before checking the theory. Check that each claim answers the question, uses the correct concept and explains why it applies. Discuss uncertain points with a teacher. The structured exercise below checks specific facts; it does not automatically mark this explanation.

Worked rule trace

A plant advisory system has two facts: SoilDry = TRUE and LeavesWilted = TRUE. Rule 1 states IF SoilDry THEN NeedsWater. Rule 2 states IF NeedsWater AND LeavesWilted THEN RecommendWatering. The inference engine applies rule 1, adds NeedsWater, then applies rule 2 to reach RecommendWatering. The explanation system can show this chain; an expert should check whether the rules are suitable.

Try new facts: SoilDry = FALSE, LeavesWilted = TRUE. Neither of these two rules establishes RecommendWatering. That does not prove watering is unnecessary; the rules provide insufficient support.