ICT Applications: Expert Systems (Chapter 6.8)
Hello IGCSE students! Welcome to one of the most interesting parts of ICT applications: Expert Systems. Don't worry if the name sounds complicated; you already use similar concepts every day. An Expert System is just a computer program designed to think and advise like a highly-trained human specialist.
In this chapter, we will learn what they are, what they are made of, and the amazing things they are used for—from diagnosing illnesses to winning chess games!
1. What is an Expert System? (Characteristics and Purpose)
An Expert System (ES) is a type of Artificial Intelligence (AI) application that mimics the decision-making ability of a human expert.
Key Characteristics of Expert Systems:
1. Specialised Knowledge: They contain a vast amount of specific knowledge about a very narrow field (e.g., only car engines, not all machines).
2. Logical Reasoning: They use logical rules and facts to solve complex problems and provide advice.
3. Explanation: Unlike many programs, an ES can usually explain its reasoning to the user, building trust and helping the user learn.
Analogy: Imagine you have a brilliant, tireless consultant who never takes a holiday. That's an Expert System!
Quick Takeaway: The purpose of an ES is to capture and apply the knowledge of human specialists to solve problems and suggest possible solutions.
2. Real-World Uses of Expert Systems (Applications)
Expert systems are essential tools across many industries where human expertise is rare, expensive, or needed 24/7.
Syllabus Examples of Expert System Uses:
* Medical Diagnosis: Assisting doctors by suggesting possible diagnoses based on patient symptoms and test results. (Example: Mycin system used to identify bacteria causing infections.)
* Car Engine Fault Diagnosis: Used by mechanics to quickly pinpoint complex faults in a vehicle's engine or electrical system.
* Mineral Prospecting: Helping geologists determine the likelihood of finding valuable resources (like oil or gold) in a specific location by analysing soil data and geological patterns.
* Financial Planning: Advising clients on investments, loans, and portfolio management based on market trends and risk profiles.
* Chess Games: Creating sophisticated AI opponents (like Deep Blue) that use complex rules and strategies to play at a master level.
* Route Scheduling for Delivery Vehicles: Optimising delivery routes for hundreds of vehicles to save time, fuel, and cost, based on traffic data and delivery windows.
* Plant and Animal Identification: Used by biologists and conservationists to identify species based on observed characteristics (e.g., leaf shape, feather colour, habitat).
Did You Know? The first widely successful expert system, DENDRAL, was developed at Stanford University in the 1960s to help chemists identify unknown organic molecules!
3. The Five Core Components of an Expert System
An expert system is not just one piece of software; it's a collection of linked components that work together. You must know these five main parts!
Mnemonic Aid: KRIES (The Expert's Toolkit)
Don't worry if this seems tricky at first—just remember the five letters: Knowledge, Rules, Inference, User, Explanation.
1. Knowledge Base
What it is: A collection of facts, data, and observations about the specific domain.
Analogy: This is the Expert's encyclopaedia or textbook.
Example: In a medical ES, the facts might be: "Headache is a symptom of flu" or "Patient X has a fever of 40°C."
2. Rules Base (or Rule Set)
What it is: A set of logical IF-THEN statements (rules) that define the relationships between the facts.
Analogy: This is the Expert's reasoning and experience.
Example: IF (Patient has fever) AND (Patient has cough), THEN (Suggest diagnosis is flu).
3. Inference Engine
What it is: The main processor. It uses the rules base to analyse the facts in the knowledge base and the input data to reach a conclusion.
Analogy: This is the Expert's brain—it performs the logical thinking.
Function: It applies search strategies (forward or backward chaining) to match symptoms to rules.
4. User Interface (UI)
What it is: The part of the system the user interacts with. It allows the user to input data and receive the results/advice.
Function: Handles the presentation of questions, scenarios, and the final solution.
5. Explanation System
What it is: A crucial part that justifies the conclusion reached by the inference engine.
Purpose: Tells the user how the system arrived at its possible solution, citing the rules it followed.
Importance: This feature is vital in areas like medicine where trust and accountability are necessary.
Quick Review: The Components
Knowledge/Rules Base: Stores all the data and logic.
Inference Engine: The 'brain' that processes the logic.
User Interface: The communication window.
Explanation System: The system's justification for its answer.
4. How an Expert System Produces Solutions
When solving a scenario, the Expert System follows a clear process of consulting its stored knowledge and applying its logical rules.
Step-by-Step Scenario Solving (e.g., Car Fault Diagnosis):
Imagine a technician uses a car diagnosis ES because a car won't start.
Step 1: Input Data (Via User Interface)
The technician inputs symptoms: "Engine turns over but won't catch," "Fuel tank is full," "Dashboard light is off."
Step 2: Inference Engine Starts Reasoning
The Inference Engine takes these inputs and searches the Rules Base.
Step 3: Applying Rules and Facts
The Inference Engine finds relevant rules:
Rule A: IF (Engine turns over) THEN (Battery is likely functional).
Rule B: IF (No Catch) AND (Fuel is present) THEN (Check ignition or spark plugs).
Rule C (From Knowledge Base): Spark plugs require electricity to fire.
Step 4: Requesting More Information
The Inference Engine may return a query to the user via the User Interface:
ES Question: "Did the technician confirm the spark plugs are producing a spark?"
Step 5: Outputting Possible Solutions
Based on all the gathered data and applied rules, the ES ranks the possible solutions.
ES Solution: Possible solution 1: Faulty ignition coil (90% certainty). Possible solution 2: Clogged fuel line (10% certainty).
Step 6: Explanation Provided
The Explanation System displays the rules and facts it used to justify the faulty ignition coil conclusion (e.g., "We concluded a faulty ignition coil because symptoms matched Rule B and Rule C, and the spark test was negative.")
Key Takeaway: The ES doesn't just give an answer; the Inference Engine systematically works through the Knowledge and Rules Base to produce a justified, possible solution.
5. Advantages and Limitations of Expert Systems
Advantages:
* Consistency: They apply the same logic every time, unlike humans who might be affected by fatigue or emotion.
* Preservation of Knowledge: They capture and store the rare knowledge of an expert, so it is not lost if the human expert retires.
* Speed: They can process complex data and reach a possible solution much faster than a human could.
* Availability: They are available 24 hours a day, 7 days a week, in any location.
Limitations:
* Lack of Common Sense: They can only operate within the bounds of their programmed knowledge and rules; they cannot handle unusual or non-standard situations.
* Knowledge Acquisition: Getting the knowledge out of the human expert and coding it into the system (the Knowledge Acquisition process) is very difficult and time-consuming.
* Narrow Focus: They are only experts in one very specific field; a medical ES cannot diagnose a car engine.
Final Tip: When discussing ES on the exam, remember that they provide possible solutions or advice, not guaranteed answers. A human expert should always review the system's output, especially in critical applications like medicine.