ComS 472: Principles of Artificial Intelligence
Department of Computer Science
Iowa State University

Fall 2009

 


STUDY GUIDE

The materials to be covered each week and the assigned readings (along with online lecture notes, if available) are included on this page. The links to lecture notes will not be in place usually until a week after the lecture. The assigned readings are divided into required and recommended readings. You will be responsible for the materials covered in the lectures and the assigned required readings. You are strongly encouraged to explore the recommended readings.


Survival Tips

  • Keep up with the assigned readings.
  • If you don't understand something in class, or the assigned problems/labs/readings, ask questions! 
  • Do not postpone working on assignments.
  • Before you come to class, review the materials from the previous lecture and come prepared to ask any questions that you might have.

Week 1 (starting August 24, 2009)

Overview of the course

Overview of artificial intelligence: What is intelligence? What is artificial intelligence (AI)? History of AI.

Introduction to intelligent agents. Rationality. PEAS description of the environment. Environment types. Structure of agents.

Required Readings -- Artificial Intelligence

Assignments

Required Readings -- Programming

You may skip most of these readings if you have prior programming experience in Java.

Recommended Materials


Week 2 (beginning August 31, 2008)

Problem-solving as state space search. Formulation of state-space search problems. Tree and Graph search algorithms. Basic search algorithms and their properties: completeness, optimality, space and time complexity. Breadth-first search, depth-first search, depth-limited and iterative deepening search.

Required readings


Week 3 (beginning September 7, 2008)

Informed (Heuristic) search. A* Search. Optimality of the A* algorithm. Heuristic functions

Required readings

Assignments


Week 4 (beginning September 14, 2009)

Local search: hill-climbing search and variants, Simulated annealing, local beam search, genetic algorithm.

Constraint satisfaction problems (CSPs): Problem solving as Constraint Satisfaction. Backtracking search, variable ordering heuristics, constraint propagation.

Required readings

Recommended readings

Assignments

  • Lab 1, due Wednesday Sep. 30, 2009, 11:00am.

Week 5 (beginning September 21, 2009)

Constraint satisfaction problems (CSPs): Local search. Problem structures.

Game-playing: The minimax algorithm. The alpha-beta-pruning algorithm. Cutting-off search for real-time decisions.

Introduction to Knowledge-based agents. Knowledge representation using logic.

Required readings

Assignments


Announcements

  • Jin Tian's office hours on Sep. 28 and 30 are cancelled.

Week 6 (beginning September 28, 2009)

Propositional Logic: Syntax and Semantics; Inference rules. Reasoning in propositional logic: the resolution rule, Horn clauses, Forward and Backward chaining. DPLL and WALKSAT algorithms.

Required readings


Week 7 (beginning October 5, 2009)

First-Order Logic (FOL): Syntax and semantics; using FOL; Knowledge engineering in FOL.

Inference in FOL: propositionalization

Required readings

Assignments


Announcements

  • The midterm is scheduled: Wednesday October 21, 2009 in class. Closed-book, closed-notes. Cover materials up to and including Chapter 9.
     

Week 8 (beginning October 12, 2009)

Inference in FOL: Unification; Forward chaining; Backward chaining; Resolution

Knowledge Representation and Reasoning Under Uncertainty. Review of elements of probability

Required readings


Week 9 (beginning October 19, 2009)

Review of elements of probability: inference, independence, Bayes rule.

Midterm exam

Bayesian Networks: Syntax and Semantics

Required readings

Assignments

Additional Information

  • Judea Pearl, Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. (1997).
  • Castillo, E., Gutierrez, J. M., Hadi, A. S. Expert Systems and Probabilistic Network Models. Berlin: Springer (1996).
  • Cowell, R. G. Lauritzen, S. L., and Spiegelhalter, D. J. Probabilistic Networks and Expert Systems Berlin: Springer (2005).
  • Korb, K.B., and Nicholson, A.E., Bayesian Artificial Intelligence, Chapman and Hall (2004).

Week 10 (beginning October 26, 2009)

Bayesian Networks: D-separation, modeling, inference.

Required readings

Assignments

  • Lab 2, due Friday November 6, 2009 11:00am.

Recommended readings

Additional Information

  • Bayesian Network software packages
  • Judea Pearl, Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. (1997).
  • Castillo, E., Gutierrez, J. M., Hadi, A. S. Expert Systems and Probabilistic Network Models. Berlin: Springer (1996).
  • Cowell, R. G. Lauritzen, S. L., and Spiegelhalter, D. J. Probabilistic Networks and Expert Systems Berlin: Springer (2005).
  • Korb, K.B., and Nicholson, A.E., Bayesian Artificial Intelligence, Chapman and Hall (2004).

Week 11 (beginning November 2, 2009)

Bayesian Networks: inference.

Learning agents. Machine learning. Bayesian decision theory: Bayes decision rule, loss function, Minimum Risk Bayes Classifier, Minimum-error-rate classification.

Parameter Estimation: Maximum-Likelihood estimation

Required readings

Recommended readings

  • Chapter 2, Duda, R., Hart, P., & Stork, D., Pattern Classification. New York: Wiley. (2001).
  • Chapter 6, Mitchell, T.,  Machine Learning. New York: Mc Graw-Hill. (1997).

Week 12 (beginning November 9, 2009)

Naive Bayes Classifier: classifying text documents. Performance evaluation: Cross-validation.

Decision tree classifier: Decision Tree Learning Algorithm (Quinlan's ID3 Algorithm)

Required readings

Assignments

  • Lab 3, due Friday November 20, 2009, 11am.

Recommended readings

  • Chapter 6, Chapter 3, Mitchell, T., Machine Learning. New York: Mc Graw-Hill. (1997).

Additional Information


Week 13 (beginning November 16, 2009)

Decision tree classifier: overfitting, missing data

Linear models for classification: Linear Discriminant Functions, Perceptrons, Multi-category classification, Winner-Take-All (WTA) networks (linear machines).

Linear models for Regression: Least Mean Squared (LMS) Error Criterion,  Gradient descent algorithm

Required readings

Recommended readings

  • Chapter 3, 4, Mitchell, T., Machine Learning. New York: Mc Graw-Hill. (1997).
  • Chapter 5, Duda, R., Hart, P., & Stork, D., Pattern Classification. New York: Wiley. (2001).

Thanksgiving break