Projects

0. Project 0: Introduction to Python (0% of grade) [No due date]
1. Project 1: Search (15% of grade) [Due Sunday September 25, 2016, 11:55pm]
2. Project 2: Constraint Satisfaction Problems (15% of grade) [Due Friday October 14, 2016, 11:55pm]
3. Project 3: Dynamic Bayesian Networks (15% of grade) [Due Sunday November 9, 11:55pm]
4. Project 4: Neural network (15% of grade) [Due Sunday December 4, 2016, 11:55pm]

Codebases for each project can be downloaded from T-square.

Homework exercises

1. Missionaries and cannibals (problem | solution)
2. Search (problem | solution)
3. Search 2 (problem | solution)
4. Rook Jumping Maze Generation. See this for more information on how the puzzles work. As a thought exercise, consider different ways to represent states using the complete-state formulation. Consider different operations for moving from possible puzzle to possible puzzle. Consider different factors that you might incorporate into an evaluation function. For example, you might want a "hard" puzzle or an "easy" puzzle; how can you write a formal definition of "easy" and "hard". Think about different algorithms that might perform well on this problem. Solution: none given: but see evaluation function designs and different possible algorithms. If you are feeling ambition, implement a puzzle generator.
5. Constraint satisfaction (problem | solution)
6. Constraint satisfaction: game of Set (problem | solution). Hint: you will have 3 variables for each member of a set, and the domain will be the different cells in the grid of cards. You will need to create some helper functions such as pattern(cell) that returns the pattern on a card in a given cell location, in addition to more common equal and not-equal operators.
7. Bayes net inference (problem | solution)
8. Bayes net inference 2 (problem | solution)
9. Markov decision process (problem | solution)
10. MDP Vvalue iteration problem from class (problem | solution)
11. Decision tree learning (problem | solution)