Neural Computing in Engineering - NUCL57500
The course presents the mathematical fundamentals of computing with neural networks and a survey of engineering applications. Computational metaphors from biological neurons serve as the basis for artificial neural networks modeling complex, non-linear and ill-posed problems. Applications emphasize the engineering utilization of neural computing to diagnostics, control, safety and decision-making problems.
Description:
The course presents the mathematical fundamentals of computing with neural networks and a survey of engineering applications. Computational metaphors from biological neurons serve as the basis for artificial neural networks modeling complex, non-linear and ill-posed problems. Applications emphasize the engineering utilization of neural computing to diagnostics, control, safety and decision-making problems.
Topics Covered:
- Basics
- Backpropagation and Related Training Algorithms
- Feedback and Other Special Neural Networks
- Dynamic Neural Networks and Control Systems
- Practical Aspects of Using Neural Networks
- Advanced Topics
Applied / Theory:
50 / 50
Homework:
A number of HW sets (8-10) will be given during the semester. These are typically problems from the book and should help with developing some computing skills
Projects:
Develop a project (preferably as a group effort) to (ideally) the level of research publication. It involves writing a report and doing a class presentation
Exams:
A midterm exam (take-home) will be given to review and sharpen your analytical skills in fuzzy mathematics
Textbooks:
Tsoukalas, L.H., Uhrig, R.E., Fuzzy and Neural Approaches in Engineering, Wiley, New York, 1997.
Hines, J.W., Matlab Supplement to Fuzzy and Neural Approaches in Engineering, Wiley, New York, 1997.