COMP3620 Artificial Intelligence
Topics in Artificial Intelligent: Design an artificial intelligent (AI) decision-making system for computer science applications. It can be used for basic sciences, mathematics, computer science, and business fundamentals. Appropriate AI algorithms (at least fuzzy logic model and artificial neural network) should be implemented and applied for linear/nonlinear system parameter training. Written report and computer codes will be an integral part of the course. Modern programming and design methodology will be stressed throughout the whole project.
1) At least a fuzzy logic model and a neural network system have to be used for programming, and two inputs and single output should be the minimum variables.
2) Programming can be undertaken in MATLAB environment or by C language or Python. The fundamental functionalities in NNs and training can be either programmed by yourself or using MATLAB toolboxes.
3) Fuzzy logic, ANFIS, LSE, gradient descent and other AI algorithms can be used for the architecture and parameter optimization or NN training.
4) Some models different from your methods can be used for performance comparison in terms of structure and parameter training.
5) It is recommended to use an error threshold of 10-4 and 200 training epochs. Project report:
1) Each person should provide a project report.
2) The report should contain sections such as Introduction, Theory Review, Model Construction, Testing and Result Analysis, Conclusion, and References, codes.
3) The project report should use 12 font size, 1.5 line space, and 1 inch page margins.
4) Report due