CSC 7991- Seminar in Neural Networks

Class information:

Course #: 7991

Credits: 3

Prerequisite: Introduction to Neural Networks

Day: Tuesday,  Thursday

Room: 0217 State Hall

Hours: 3.00-4.20pm

Instructor information:

Instructor: Dr. Sorin Draghici

Office: 420 State Hall

Office hours: Tuesday 6pm-7.00pm, Thursday 6.00pm-7.00pm or by appointment

Telephone: 577-5484


Web page:

On this web page you can find the syllabus, the transparencies used during the course and announcements regarding the course if any.


Neural Network Design - Martin T. Hagan, Howard B. Demuth, Mark Beale, 1995
Self-Organizing Maps - Teuvo Kohonen, Second Edition, 1997
Circuit Complexity and Neural Networks - Ian Parberry, MIT Press, 1994
The course is intended for senior undergraduate and graduate students interested in neural networks. The course may appeal to students with a background in computer science, mathematics, physics or any engineering field or to anybody interested in this particular topic.
Neural networks are inspired by the way our brain works and they are one of the most powerful machine learning techiques known to man to date. Neural network are able to learn from examples without the need for an explicit understanding of the phenomenon and without any programming. They can also generalize their knowledge to exhibit proper behaviour in unseen cases, very similarly to the way humans do. The range of applications for neural networks includes pattern recognition, control, classsification, function approximation, optimization, etc. The most advanced artifacts of modern technology today -from  electric razors to Ford automobiles- include neural networks technology. This course is the follow-up for "Introduction to Neural Networks".


CSC785 - "Introduction to Neural Networks".
The students will acquire specific knowledge about the field as well as general research skills. The lectures will continue to present the most important neural network paradigms and techniques with an emphasis to solving real world problems. The course will discuss basic principles of research in general and research in neural networks in particular. During the course, the students will be required to undertake a project. The report for this project will be the most important component of the grade and will be written in the form of a research paper suitable for a technical publication. The actual submission for publication of such paper is encouraged but not required.
Course contents - Class plan
In the first part, the course will continue to cover several chapters from the indicated textbook. Techniques not included in the textbook will be discussed. In the second part of the course, students will present the results of their project work. This project work will include reading research papers and presenting them in class.
The instructor reserves the right to vary the level and depth of the material covered in order to adapt the course to the background and level of the students. Furthermore, some topics may be added if time permits.
Class policies

Attendance: Attending all lectures is essential; the assignments, exams, etc. will be based primarily (though not exclusively) on the materials presented in these lectures. Also, assignments due dates, explanation and clarification of assignments and material outside the textbooks will be presented during lecture sessions. If you miss a lecture, it is your responsibility to obtain the information covered in the session.

Health Safety: Please report to the instructor any health condition which may create a classroom emergency (e.g. seizure disorders, diabetes, heart conditions, etc.).

Computer lab: To enhance your learning and for your homework, the computer lab, equipped with PC's and Unix workstations is available to you during the time posted on the lab's door.

Grading procedures

The grading will include a project  in which the students will demonstrate their understanding of the techniques presented, a classroom presentation  and a final exam. The report for the project will be written as a research paper suitable for  publication in a technical journal or conference.

The final grade will be calculated as follows:

Average of

Project: 60% 

Class discussion: 25% 

Final exam: 15%

The final letter grade will be determined approximately as follows: 

A+: 95-100 % 
A: 90-94.99 
A-: 85-89.99 
B:+ 80-84.99 
B: 75-79.99 
B-: 70-74.99 
C+: 65-69.99 
C: 60-64.99 
F: less than 60% 

A grade of Incomplete (I) will not be given unless in very exceptional circumstances. 

Student Responsibilities:

As a college student who is committed to seek a higher education, we expect you be a very responsible person. At the least, please:

Important dates: