(50 points) For your class project, you need to implement an
AI algorithm. Several ideas are given below. Choose one of these ideas
or create one of your own, but you need to turn in your project proposal
for approval by November 13. If this proposal is not submitted, two points
will be deducted from your project grade.
With the project, turn in your code, sample output, a description of how to
compile and run the project, and a brief description (less than one page) of
the project and its design. You can write this program in any language as
long as it can run on the class machine. Make sure that a majority of your
code is your own, not downloaded from a remote site. If you use any materials
from books or the web, reference that material. Two-person projects may be
considered, although the distribution of effort between the individuals must
be made clear in the project proposal and in the final project report.
I will be grading the project based on the implementation completeness
and correctness, any enhancements you add to the basic idea, the thoroughness
of experimentation or analysis, and the report. I will also factor in the
difficulty of the project you undertake. Be careful, though - it is better
to do a good job on a smaller project than to turn in an incomplete larger
project.
Project ideas
Enhance the eight puzzle program by adding RBFS search or another
more recent search technique.
Use bidirectional search to find a path of links between two input URLs.
Implement minimax with alpha-beta pruning and secondary search for
the game of Othello.
Extend basic minimax and alpha-beta pruning to handle three-player games.
Write a theorem prover in Prolog for proof by refutation.
Design a belief network for a particular problem where the structure and
weights are learned from sample data.
Use a recurrent neural network to recognize gestures from a sequence
of images (a short video).
Implement a two-player game that involves chance (e.g., backgammon).
Implement a reinforcement learning agent that plays the Wumpus World
game (a C++ Wumpus World simulator is available at
http://eecs.wsu.edu/~holder/courses/cse5361/wumpus.html)
Compare logic-based and graph-based approaches to relational learning
applied to problems in computational chemistry.
See http://www.eecs.wsu.edu/~holder/aiproject for details.
