Artificial intelligence

1. Overview

Artificial intelligence (AI) is transforming the world around us, from the apps on our phones to how businesses operate. Understanding AI is crucial for computer scientists because it is a rapidly growing field that presents exciting opportunities and ethical considerations.

Key Definitions

• Artificial Intelligence (AI): Computer systems that can perform tasks that normally require human intelligence, such as learning, problem-solving, decision-making, understanding language, and recognizing patterns.
• Machine Learning (ML): A type of AI where systems learn from data without being explicitly programmed.
• Expert System: An AI system designed to mimic the decision-making abilities of a human expert in a specific domain.
• Knowledge Base: A database of facts and information about a specific subject or domain used by an expert system.
• Rule Base: A set of IF-THEN rules based on expert knowledge used by an expert system to make decisions.
• Inference Engine: The part of an expert system that applies the rules in the rule base to the facts in the knowledge base to draw conclusions.

Core Content

Understanding Artificial Intelligence (AI)

• AI aims to create machines that can think and act like humans. This includes tasks like:

  • Learning: Improving performance over time based on data.
  • Problem-solving: Finding solutions to complex issues.
  • Decision-making: Choosing the best option from a set of possibilities.
  • Understanding language: Interpreting and responding to human language.
  • Recognizing patterns: Identifying trends in data.

• Examples of AI in action:

  • Voice assistants: Siri, Alexa, Google Assistant.
  • Recommendation systems: Netflix, Amazon product suggestions.
  • Image recognition: Identifying objects in photos, facial recognition.
  • Game playing: AI that can play games like chess or Go at a high level.

Characteristics of AI

• Collection of Data: AI systems require large datasets to analyze and learn from. The more data, the better the AI can perform.
• Rules for Using Data: Algorithms define how the AI system should process the data. These algorithms are often complex mathematical models.
• Ability to Reason: AI can draw logical conclusions based on the data it has been given and the rules it follows.
• Ability to Learn (Machine Learning): The AI system can improve its performance over time by analyzing new data and adjusting its algorithms. This is often called "training".
• Ability to Adapt: The AI system can change its behaviour based on new situations or changing data.

Basic Operation and Components of AI Systems

• Expert Systems:

  • Mimic the decision-making process of a human expert.

  • Useful in fields where expertise is scarce or expensive (e.g., medical diagnosis).

  • Components:

    • Knowledge Base: Contains facts about the specific area of expertise.

    • Rule Base: Contains IF-THEN rules that represent the expert's knowledge. Example: IF patient has fever AND patient has cough THEN suspect influenza

    • Inference Engine: Applies the rules in the rule base to the facts in the knowledge base to draw conclusions. It essentially "reasons" using the knowledge and rules.

    • Interface: Allows users to interact with the system, ask questions, and receive advice.

*   **Example: Medical Diagnosis System**

    *   **Knowledge Base:** Facts about diseases, symptoms, and medical history.
    *   **Rule Base:** Rules linking symptoms to possible diagnoses.
    *   **Inference Engine:** Analyzes patient data and applies rules to suggest possible diagnoses.
    *   **Interface:** Allows doctors to input patient information and view the system's recommendations.

• Machine Learning:

  • Systems learn from data without explicit programming.

  • Algorithms identify patterns and relationships in the data.

  • The more data the system is trained on, the better it performs.

  • Example: Image Recognition

    • A machine learning system is trained on a large dataset of images labeled with the objects they contain (e.g., cats, dogs, cars).

    • The system learns to identify the features that distinguish each object.

    • When presented with a new image, the system can identify the objects it contains based on the features it has learned.

*   **Example: Spam Filtering**

    *   A machine learning system is trained on a large dataset of emails labeled as either "spam" or "not spam".
    *   The system learns to identify the features that are common in spam emails (e.g., certain words, suspicious links).
    *   When a new email arrives, the system analyzes its features and predicts whether it is spam or not.

Exam Focus

• AI Definition: Examiners want a clear understanding of what AI is and what it can do. Include aspects like learning, problem-solving, and decision-making.
• AI Characteristics: Be able to explain how data collection, rules, reasoning, learning, and adaptation contribute to AI systems.
• Expert Systems: Understand the four key components (knowledge base, rule base, inference engine, interface) and how they work together. Be able to give examples.
• Machine Learning: Emphasize the importance of large datasets for training and the ability of systems to learn from data without explicit programming. Relate to examples like image recognition or spam filtering.
• Contextual Application: Examiners often assess your ability to apply AI concepts to real-world scenarios. Make sure you understand examples of AI in different contexts.
• Technical Terminology: Use precise terms like "knowledge base," "inference engine," "machine learning," and "algorithm."

Common Mistakes to Avoid

• ❌ Wrong: "AI is just robots." ✓ Right: "AI is a computer system that can perform tasks that usually need human intelligence, like making decisions."
• ❌ Wrong: "Expert systems don't need data." ✓ Right: "Expert systems use a knowledge base, which is a collection of facts and information about a specific topic."
• ❌ Wrong: "Machine learning means programming everything." ✓ Right: "Machine learning means the system learns from data without being specifically programmed for every situation."
• ❌ Wrong: Giving vague examples without explanation. ✓ Right: Providing specific examples (e.g., medical diagnosis, spam filtering) and explaining how AI is used in those situations.
• ❌ Wrong: Only defining AI and not explaining its characteristics or how it works. ✓ Right: Provide an overview and then expand into AI characteristics and operational components.

Exam Tips

• Understand the Context: Read the question carefully and identify the specific context (e.g., medical diagnosis, spam filtering).
• Use Keywords: Include key terms like "knowledge base," "rule base," "inference engine," and "machine learning" in your answers.
• Provide Examples: Back up your explanations with real-world examples of AI applications.
• Explain the Process: Don't just state facts; explain how the AI system works, the steps involved in learning, or the decision-making process.