CSC403: Artificial Intelligence Lecture 1: Introduction

Course Overview

  • Course Name: CSC403: Artificial Intelligence

  • Textbook: Artificial Intelligence: A Modern Approach, 4th edition

Necessary Skills

  • Essential: Programming, curiosity, enthusiasm, dedication

  • Helpful Skills: Probability and statistics, linear algebra, discrete math, calculus, logic

Course Work

  • Quizzes: 5%

  • Homework Assignments: 5%

  • Lab Exam/Project: 10%

  • Midterms: 40% (2 exams)

  • Final Exam: 40%

  • Late Penalties: Graded, with 3 "free late days"

Main Topics in AI

  1. Introduction to AI

  2. Uninformed Search Methods

  3. Informed Search Methods

  4. Adversarial Search

  5. Expectimax Search

  6. Constraint Satisfaction Problems

  7. Markov Decision Process

  8. Reinforcement Learning

  9. Logical Agents

  10. Swarm Intelligence

Definitions of AI

  • David Marr: Study of information processing problems from biological roots.

  • Rodney Brooks: Intelligent connection of perception to action.

  • Alan Turing: Actions indistinguishable from human actions.

Perspectives in AI

  • Categories:

    • Thinking humanly

    • Thinking rationally

    • Acting humanly

    • Acting rationally

  • Focus: Algorithms for rational thinking/acting.

Turing Test

  • Evaluates if machines can exhibit intelligent behavior.

  • Major components: Knowledge representation, reasoning, natural language processing, learning.

Challenges of AI

  • Involves complex data (speech, images, etc.) and often noisy.

  • Issues with defining competence for tasks and developing efficient algorithms.

Task Characteristics

  • Observability: Fully vs. partially observable.

  • Determinism: Deterministic vs. stochastic tasks.

  • Episodic vs. Sequential: Independence of actions.

  • Static vs. Dynamic: Environment changes during tasks.

  • Discrete vs. Continuous: Number of states.

  • Single vs. Multi-agent systems.

Properties of Intelligence

  • Interaction with real world, representation, learning, reasoning, and decision making.

AI Paradigm

  • Develop general methods for solving real-world tasks by modeling and algorithmic problem solving.

AI Applications

  • Numerous, including language translation, speech/image recognition, autonomous vehicles, etc.

A Brief History of AI

  • 1943-1950: Early models and concepts.

  • 1950s: Dartmouth Conference and early programming.

  • 1970-90: Knowledge-based systems boom and bust.

  • 1990-Present: Statistical approaches; resurgence with neural networks and big data.