This course examines the use of machine learning techniques in all stages of game design. Topics covered include environment and character modeling, motion synthesis, behavior learning, evolution and competition. The emphasis will be on cutting edge technology that utilizes the vast amounts of recorded game data as a basis for learning more realistic or more effective game design strategies. Advanced topics that will also be covered are gamebot identification in online games, as well as integration and evaluation of learning in games. Students will participate in groups to develop a game using principles learned in class. To complete the project, they will be required to implement and observe a representative set of the techniques covered in class.

Goals: Students will acquire a clear understanding of machine learning concepts related to game design, as well as hands-on experience in the different aspects of designing a complete game.

 Prerequisites: MA 222, CS585 (Game Design), CS539 or CS587 (Game Engine Design), or instructor approval

Course Outcomes   

Each course outcome is followed in parentheses by the Program Outcome to which it relates.

1.  Explain and implement procedural modeling in 3D.  [problem-solving]

2. Create virtual creatures using standard graphics software. [languages, problem-solving]

3.  Explain motion synthesis from exemplars. [problem-solving]

4. Implement natural evolution in virtual worlds. [languages, problem-solving]

5. Implement reactive behavior learning using simple neural networks. [languages, problem-solving]

6.  Explain and use the LEE and Isomap algorithms for nonlinear dimensionality reduction.  [problem-solving]

7. Implement of gamebot detection from real game traces recorded online using manifold learning techniques. [languages, problem-solving]

8. Evaluate effectiveness of learning on their virtual character by competing against classmates in virtual platform. [requirements, problem-solving]

9. Integrate software modules to produce final game prototype. [languages]

Required Texts:  Research papers  see  WeBCT  to download

Syllabus

Final Grade:

HW 50%

Project design documentation & presentation 10%

Final Project: 40%

Stevens Institute, Department of Computer Science
Last Update: 01/24/10