Machine learning vs rule-based systems in AI for Everyone - Hands-On Comparison

Machine learning vs rule-based systems

📖 Scenario: You are working on a simple AI assistant that can decide how to respond to user questions. You want to understand the difference between machine learning and rule-based systems by creating examples of each.

🎯 Goal: Build two simple examples: one using a rule-based system with fixed rules, and one using a machine learning style approach with data and a simple decision logic. This will help you see how each system works in practice.

📋 What You'll Learn

Create a dictionary called rules with fixed question-answer pairs

Create a list called training_data with example questions and answers

Write a function called rule_based_response that uses the rules dictionary to find answers

Write a function called ml_style_response that uses simple matching on training_data to find answers

💡 Why This Matters

🌍 Real World

Understanding the difference between rule-based and machine learning systems helps when designing AI assistants, chatbots, or decision-making software.

💼 Career

Many AI and software development jobs require knowledge of how to implement and choose between rule-based logic and machine learning models.

1

DATA SETUP: Create the rule-based system dictionary

Create a dictionary called rules with these exact entries: 'hello': 'Hi there!', 'bye': 'Goodbye!', and 'thanks': 'You are welcome!'.

AI for Everyone

# Create the rules dictionary with fixed question-answer pairs
# Your code here

Hint

Use curly braces {} to create a dictionary with keys and values.

2

CONFIGURATION: Create the machine learning style training data

Create a list called training_data with these exact tuples: ('hello', 'Hi there!'), ('bye', 'Goodbye!'), and ('thanks', 'You are welcome!').

AI for Everyone

rules = {
    'hello': 'Hi there!',
    'bye': 'Goodbye!',
    'thanks': 'You are welcome!'
}
# Create the training_data list with example question-answer pairs
# Your code here

Hint

Use square brackets [] to create a list of tuples.

3

CORE LOGIC: Write the rule-based response function

Write a function called rule_based_response that takes a parameter question and returns the answer from the rules dictionary if the question exists, or 'I do not understand.' if it does not.

AI for Everyone

rules = {
    'hello': 'Hi there!',
    'bye': 'Goodbye!',
    'thanks': 'You are welcome!'
}

training_data = [
    ('hello', 'Hi there!'),
    ('bye', 'Goodbye!'),
    ('thanks', 'You are welcome!')
]

# Write the rule_based_response function below
# Your code here

Hint

Use an if statement to check if the question is a key in the rules dictionary.

4

COMPLETION: Write the machine learning style response function

Write a function called ml_style_response that takes a parameter question and searches the training_data list for a matching question. If found, return the answer; otherwise, return 'I do not understand.'. Use a for loop with variables q and a to iterate over training_data.

AI for Everyone

rules = {
    'hello': 'Hi there!',
    'bye': 'Goodbye!',
    'thanks': 'You are welcome!'
}

training_data = [
    ('hello', 'Hi there!'),
    ('bye', 'Goodbye!'),
    ('thanks', 'You are welcome!')
]

def rule_based_response(question):
    if question in rules:
        return rules[question]
    else:
        return 'I do not understand.'

# Write the ml_style_response function below
# Your code here

Hint

Use a for loop to check each tuple in training_data for a matching question.