Artificial Intelligence Course

Free online course · 10 lessons

Learn the core methods of artificial intelligence: search, constraint satisfaction, games, logic and planning, probabilistic reasoning, Markov models, decision making and reinforcement learning, machine learning and neural networks, and the ethics and governance of AI. After the course you can build simple intelligent agents and judge AI systems critically.

By Studeli

Start with lesson 1

What you will learn in 10 lessons

Every lesson includes all nine study tools

  • Study Guide
  • Summary
  • Walkthroughs
  • Practice Questions
  • Flashcards
  • Explain It
  • Key Terms
  • Cheat Sheet
  • Concept Map

  1. Intelligent Agents

    Start here

    Describe what AI is, model problems as agents acting in environments and formulate state spaces, actions and goals.

    5 learning goals
    1. 1.Explain what artificial intelligence is and how the field developed
    2. 2.Describe agents, environments and rational behavior
    3. 3.Classify task environments
    4. 4.Compare agent architectures
    5. 5.Formulate problems as state-space search

  2. Uninformed and Informed Search

    Solve problems with breadth-first, uniform-cost and A* search, design admissible heuristics and use local search for very large spaces.

    5 learning goals
    1. 1.Frame problems as search and judge search algorithms
    2. 2.Search without domain knowledge
    3. 3.Search with heuristics: greedy best-first search and A*
    4. 4.Design and evaluate heuristics
    5. 5.Use local search for very large spaces

  3. Constraint Satisfaction Problems

    Formulate scheduling and assignment problems as constraint satisfaction problems and solve them with backtracking, constraint propagation and ordering heuristics.

    5 learning goals
    1. 1.Formulate problems as constraint satisfaction problems
    2. 2.Solve CSPs with backtracking search
    3. 3.Prune domains with constraint propagation
    4. 4.Solve CSPs with local search
    5. 5.Exploit the structure of a problem

  4. Adversarial Search

    Choose moves in games with minimax, alpha-beta pruning and expectimax, and design evaluation functions.

    6 learning goals
    1. 1.Describe games as search problems and draw their game trees
    2. 2.Choose optimal moves with minimax
    3. 3.Speed up minimax with alpha-beta pruning
    4. 4.Play well under time limits with evaluation functions
    5. 5.Choose moves when chance is involved
    6. 6.Search large games with Monte Carlo tree search

  5. Logical Agents and Planning

    Represent knowledge in propositional and first-order logic, draw inferences and formulate classical planning problems.

    6 learning goals
    1. 1.Represent what an agent knows in propositional logic
    2. 2.Prove conclusions with propositional inference and SAT solvers
    3. 3.Represent knowledge in first-order logic
    4. 4.Draw inferences in first-order logic
    5. 5.Organize knowledge in ontologies and knowledge graphs
    6. 6.Formulate and solve classical planning problems

  6. Bayesian Networks

    Reason under uncertainty with probability and Bayes' rule, represent dependencies in Bayesian networks and compute answers with exact and sampling-based inference.

    5 learning goals
    1. 1.Reason with probability, Bayes' rule and independence
    2. 2.Represent dependencies with Bayesian networks
    3. 3.Compute exact answers with enumeration and variable elimination
    4. 4.Approximate answers by sampling
    5. 5.Make decisions with utilities, decision networks and the value of information

  7. Hidden Markov Models

    Model changing systems with Markov chains and hidden Markov models, and track hidden states with filtering and particle filters.

    5 learning goals
    1. 1.Model a changing system with a Markov chain
    2. 2.Represent hidden states with hidden Markov models
    3. 3.Infer hidden states from a sequence of evidence
    4. 4.Approximate filtering with particle filters
    5. 5.Apply temporal models to tracking and speech

  8. Reinforcement Learning

    Model sequential decisions as Markov decision processes, solve them with value and policy iteration and learn policies from experience with Q-learning.

    5 learning goals
    1. 1.Model sequential decisions as Markov decision processes
    2. 2.Express values with Bellman equations
    3. 3.Solve known MDPs with dynamic programming
    4. 4.Learn about an unknown MDP from experience
    5. 5.Learn to act with Q-learning and exploration

  9. Supervised Learning and Neural Networks

    Train simple classifiers such as naive Bayes, perceptrons and logistic regression, and explain how neural networks learn with gradient descent.

    6 learning goals
    1. 1.Set up a supervised learning problem
    2. 2.Classify with naive Bayes
    3. 3.Train linear classifiers
    4. 4.Optimize models with gradient descent
    5. 5.Build and train neural networks
    6. 6.Recognize and control overfitting

  10. AI Ethics

    Evaluate AI systems for fairness, transparency, privacy and safety risks, and explain how regulation and governance frameworks manage those risks.

    6 learning goals
    1. 1.Analyze the ethics of an AI system
    2. 2.Detect, measure and reduce unfair bias
    3. 3.Make AI systems transparent and protect privacy
    4. 4.Make AI systems safe, robust and aligned
    5. 5.Assess the societal and labor impacts of AI
    6. 6.Explain how regulation and governance frameworks manage AI risk

What to take next

More Computer Science courses
Introduction to Computer Science

Make your own Study Materials

Turn your notes, slides and documents into Study Guides, Practice Questions, Flashcards and 6 more tools.

Get Started
Start with lesson 1