Information Technology (9626) | Expert systems

Study Notes: Expert Systems (AS Level Topic 7)

Welcome to Topic 7! This chapter explores Expert Systems, one of the most exciting applications of Information Technology.
These systems are essentially computer programs designed to act like human specialists or consultants. Understanding them is key to seeing how IT can solve highly complex, real-world problems. Don't worry if the terminology seems complex—we'll break down the components one by one!

7.1 Expert Systems: The Digital Specialist

What is an Expert System?

An expert system is an application program that simulates the judgment and behaviour of a human expert in a specific, narrow domain.
They are designed to produce possible solutions or advice for different scenarios, just like a human consultant would.

Analogy: Imagine an expert system is a highly trained detective. It has a massive book of rules (knowledge) and uses logic (the inference engine) to solve a mystery (the problem).

How Expert Systems are Used (Scenarios)

Expert systems excel in areas that require high levels of specialised knowledge and consistent decision-making. The syllabus requires you to know several key applications:

• Medical Diagnosis: Identifying potential illnesses based on patient symptoms and test results.
• Car Engine Fault Diagnosis: Pinpointing mechanical issues based on vehicle sensor readings and described symptoms.
• Financial Planning & Investment Analysis: Advising clients on stocks, savings, and retirement plans based on market data and risk profiles.
• Insurance Planning: Calculating risks and appropriate premiums for complex policies.
• Mineral Prospecting: Analysing geological survey data to determine the best locations for mining valuable minerals.
• Route Scheduling for Delivery Vehicles: Calculating the most efficient and cost-effective routes for a fleet of vehicles (like courier services).
• Plant and Animal Identification: Using characteristics (leaf shape, habitat, etc.) to identify species.

Key Takeaway: Expert systems are used where speed, consistency, and highly specialized knowledge are needed to solve complex, non-mathematical problems.

The Five Core Components of an Expert System

An expert system needs several crucial parts to function effectively. You must know all five:

1. The Knowledge Base

This is the heart of the system—it holds all the information the expert knows. It contains two types of content:

• Facts: Basic information (e.g., "A fever is a high body temperature.").
• Rules Base: The set of logical procedures, usually written as IF...THEN statements (e.g., "IF the patient has a cough AND a fever, THEN consider influenza.").
2. The Inference Engine

This is the "brain" that processes the information. The inference engine searches the knowledge base, uses the rules, and applies logic to draw conclusions or ask further questions. It determines *how* to apply the IF...THEN rules.

3. The User Interface (UI)

This is how the non-expert user interacts with the system. It asks questions, accepts user inputs (data/symptoms), and presents the final advice or diagnosis. It needs to be clear and easy to use.

4. The Explanation System

This component is vital for building user trust. It can explain *how* the expert system reached a particular conclusion or why it asked a certain question. For example, a medical diagnosis system can state: "The system concludes 'Influenza' because the patient has a fever (Fact 1) and a cough (Fact 2), which matches Rule 5 in the Knowledge Base."

5. The Knowledge Base Editor

This tool is used by the human expert (or knowledge engineer) to update, refine, and maintain the facts and rules within the Knowledge Base. This ensures the system stays current with new research or emerging threats (like new diseases or car models).

Chaining: How the Inference Engine Reasons

The inference engine uses two main methods, or "chains," of reasoning to process rules and find solutions. These chains rely heavily on the IF...THEN constructs stored in the Knowledge Base.

1. Forward Chaining (Data Driven)

Process: The system starts with the available data (facts or inputs) and works forward, applying rules to see what conclusions can be drawn.

Step-by-Step Example (The Data drives the conclusion):

1. Input Data: "Temperature is high."
2. Rule 1: IF temperature is high THEN status is 'fever'.
3. Rule 2: IF status is 'fever' THEN ask for headache.
4. Conclusion: The system now knows the patient has a fever and asks about a headache.

When is it used? When the input data is known, and the goal is to see all possible outcomes. This is often called Data Driven reasoning.

• Uses: Gaming (determining outcomes based on player actions), Artificial Intelligence (for use in manipulating social media—reacting to specific keywords or behaviours).

2. Backward Chaining (Goal Driven)

Process: The system starts with a specific goal or hypothesis and works backward to see if the required data (facts) supports that goal. It asks questions to verify the supporting facts.

Step-by-Step Example (The Goal drives the questioning):

1. Goal/Hypothesis: "Patient has disease X."
2. Rule: To confirm disease X, patient must have symptom A and symptom B.
3. System checks: Does the patient have symptom A? (If no, discard hypothesis X. If yes, proceed.)
4. System asks the user: "Does the patient have symptom B?"

When is it used? When there is a specific target diagnosis or solution. This is often called Goal Driven reasoning.

• Uses: Most forms of Diagnoses (medical, mechanical, system fault), where the system works backwards from a potential problem.

Advantages and Disadvantages of Expert Systems

Like any technology, Expert Systems offer great benefits but also have limitations. Exam questions often ask you to evaluate these points.

Advantages of Expert Systems

• Consistency: They provide the same quality of advice every time, unlike human experts whose judgment can be affected by fatigue or mood.
• Speed: They can process large amounts of data and rules much faster than a human expert, leading to quicker diagnoses or analyses.
• Permanent Knowledge: The knowledge is stored and doesn't retire or die. It can be easily duplicated and distributed to different locations.
• Hazardous Environments: They can be used to provide advice in dangerous situations (e.g., fault diagnosis in a nuclear reactor) where human presence is unsafe.
• Training Tool: They can be used to train junior staff by showing them the reasoning (via the Explanation System) used by the simulated expert.

Disadvantages of Expert Systems

• Cost and Maintenance: Developing and maintaining the Knowledge Base and the software is very expensive and time-consuming.
• Lack of Intuition: Expert systems lack common sense, creativity, or "gut feeling" (intuition). They can only operate based on the rules explicitly given to them.
• Narrow Focus: They are typically limited to a very specific domain and cannot apply knowledge outside that area. (A medical ES cannot diagnose a car fault).
• Knowledge Acquisition Difficulty: Extracting the knowledge (the rules) from a human expert and structuring it into IF...THEN format is often a major challenge.
• Mistakes in Knowledge Base: If the rules in the Knowledge Base are wrong or incomplete, the resulting advice will also be wrong.

Did you know? The process of gathering knowledge from a human expert and converting it into a structured format for the Knowledge Base is called Knowledge Engineering. This is often the hardest and most time-consuming part of creating an expert system.

Key Takeaway and Revision Checklist

Expert systems combine a Knowledge Base (facts and IF...THEN rules) with an Inference Engine to solve problems. They use Forward Chaining (Data Driven) or Backward Chaining (Goal Driven) depending on whether you start with the data or the desired conclusion. Remember their consistency and speed are major advantages, but their lack of human intuition is a critical weakness.