Expert systems

Welcome to the World of Expert Systems!

In this chapter, we are going to explore Expert Systems. Imagine you have a friend who knows absolutely everything about car engines or a doctor who has read every medical textbook ever written and never forgets a single detail. An expert system is exactly that—a computer program designed to act like a human expert to solve complex problems.

Don't worry if this sounds a bit futuristic! By the end of these notes, you’ll see that they are just logical tools used to help humans make better decisions. Let’s dive in!

1. What is an Expert System?

An Expert System is a piece of software that uses Artificial Intelligence (AI) to simulate the judgment and behavior of a human or an organization that has expert knowledge and experience in a particular field.

Analogy: Think of an expert system like a "digital consultant." You give it facts about a situation, and it uses its "brain" (the software) to give you advice or a solution.

2. The Components of an Expert System

To work correctly, an expert system needs several parts working together. You can think of these as the different parts of a human expert's mind.

The User Interface

This is how the user interacts with the system. It usually involves a screen where the system asks questions (like "Does the patient have a fever?") and the user types in answers.

The Knowledge Base

This is the "library" of the system. It contains two things:
1. Facts: Basic truths about the subject (e.g., "Water boils at 100°C").
2. Rules Base: A set of IF...THEN constructs (e.g., "IF the engine won't start AND the lights are dim, THEN the battery is likely dead").

The Inference Engine

This is the "engine" or the "brain" that does the thinking. It looks at the facts provided by the user and searches through the Knowledge Base to find a match using the Rules Base. It links the user's answers to the rules to come up with a conclusion.

The Explanation System

A human expert can tell you why they gave you a certain piece of advice. The Explanation System does the same thing. It shows the user the logic it used to reach its conclusion, which helps build trust in the system.

The Knowledge Base Editor

Knowledge changes over time (for example, new medicines are invented). The Knowledge Base Editor allows a human expert to update the facts and rules in the system without needing to rewrite the whole program.

Quick Review:
• User Interface: The screen/input method.
• Knowledge Base: The facts and rules.
• Inference Engine: The logic/thinking part.
• Explanation System: Tells you "why."
• Knowledge Base Editor: For updating information.

3. Real-World Scenarios

Expert systems aren't used for everything—they are used for specific, high-stakes tasks. Here are the common scenarios you need to know for your exam:

• Medical Diagnosis: Helping doctors identify diseases based on symptoms.
• Car Engine Fault Diagnosis: Helping mechanics find out why a car isn't running properly.
• Mineral Prospecting: Analyzing soil and rock data to predict where gold or oil might be hidden underground.
• Investment Analysis: Predicting which stocks or bonds will make the most money.
• Financial Planning: Helping people decide how to save for retirement.
• Insurance Planning: Determining the best insurance policy for a customer's specific risks.
• Route Scheduling for Delivery Vehicles: Finding the fastest way for a fleet of trucks to deliver packages.
• Plant and Animal Identification: Using features (like leaf shape or fur type) to name a species.

4. How Expert Systems "Think": Chaining

The Inference Engine uses two main methods to arrive at a solution. This is often where students get a bit confused, but it’s simpler than it looks!

Forward Chaining (Data Driven)

This starts with the available data and uses the Rules Base to see where that data leads. It is "data driven."

Example: You tell a medical system "I have a cough" and "I have a fever." The system looks at these facts and moves forward to conclude: "You have the flu."

Backward Chaining (Goal Driven)

This starts with a goal or a conclusion and works backward to see if the facts support it. It is "goal driven."

Example: A system starts with the goal: "Does this patient have the flu?" It then looks backward to ask: "Does the patient have a fever?" and "Does the patient have a cough?" If the facts are there, the goal is confirmed.

Memory Trick:
• Forward starts with Facts.
• Backward starts with the Big goal.

5. Data Driven vs. Goal Driven Applications

We use these different types of logic in different areas:

• Diagnoses: Usually use Forward Chaining (starting with symptoms to find a disease).
• Gaming: Non-player characters (NPCs) often use Goal Driven logic. Example: An NPC has the goal to "Attack the player." It then checks if it has a weapon and if the player is in range.
• Artificial Intelligence in Social Media: Used to manipulate what you see. The system has a Goal (to keep you on the app longer) and works backward to decide which posts will make that happen.

6. Advantages and Disadvantages

Like any technology, expert systems have their pros and cons.

Advantages

• 24/7 Availability: Computers don't need to sleep, eat, or take holidays.
• Consistency: Unlike humans, the system won't have a "bad day" and give a different answer to the same problem.
• Multiple Experts: A system can combine the knowledge of many different experts into one place.
• Use in Dangerous Areas: They can be used in places where it might be unsafe for a human expert (like a nuclear plant).

Disadvantages

• No "Common Sense": A system follows rules blindly. It can't tell if a user provides an answer that is obviously ridiculous.
• High Costs: They are very expensive and time-consuming to build and maintain.
• Garbage In, Garbage Out: If the rules or facts in the Knowledge Base are wrong, the advice will be wrong.
• Lack of Emotion/Empathy: In fields like medicine, a computer can't provide the "human touch" or emotional support a real doctor can.

Key Takeaway: Expert systems are powerful because they are consistent and never forget, but they are limited by the quality of the rules humans give them and their lack of general human intuition.

Common Mistakes to Avoid

• Mistake: Thinking the Inference Engine is the Knowledge Base.
• Correction: The Knowledge Base is the storage (books); the Inference Engine is the processor (the person reading the books).
• Mistake: Thinking Expert Systems are "creative."
• Correction: They are strictly logical. They only know what they have been told via the IF...THEN rules.

Don't worry if the logic of chaining feels a bit abstract at first. Just remember that Forward starts with what you know, and Backward starts with what you want to prove!