The semester project for this course will consists of a working program written in any language of the student's choosing that demonstrates the principles of Artificial Intelligence taught in this course. The program will be accompanied by a descriptive paper of 5-10 pages that explains and analyzes the operation of the program.
All projects will be evaluated using two criteria: the sophistication of the AI techniques in the program and the quality of the analysis in the paper.
Examples of acceptable projects are:
1) An agent that plays a game. For example, in previous semesters, some students have created agents for 3D tic-tac-toe. A game-playing agent must demonstrate some novel aspects, e.g. it cannot just use minimax but must employ one or more new heuristics devised by the student.
2) An agent that performs a useful task. For example, some students have programmed agents that find the best route in the NYC subway system or for a taxi in Manhattan. Such an agent cannot just employ a simple search method such as depth-first iterative deepening, but must combine methods to achieve good performance.
3) A comparison of two or more distinct search methods on a task. Almost any task can be chosen. The emphasis in this type of project is on the analysis in the paper, which must include good statistics on the runtime (in terms of the number of nodes expanded) of the methods, and an explanation of the results.