Neural Networks and Fuzzy Systems

Course No: CSE 4203 

Course title: Neural Networks and Fuzzy Systems

Prerequisite courses: N/A

Contact hours/week: 3

Credits: 3.00

Course Rationale:

This course aims at introducing the fundamental theory and concepts of computational intelligence methods, in particular neural networks, fuzzy systems, and genetic algorithms, and their applications in the area of machine intelligence. This will help students to get sufficient knowledge to analyze and design the various intelligent control systems in real world scenarios.

Course Content / Course Description:

Introduction to the Concept: Human Brain Mechanisms and Neural Machine Intelligence.

Fundamental Neural Network Concepts: Artificial Neuron Basic Models, Activation Function, Network Architecture, Neural Network Viewed as a Directed Graph  Basic Learning  Rules,  Perceptrons Overview, Single Layer Perceptrons, Single Layer Perceptrons Mathematical Model, Perceptrons Learning Algorithm, Delta Learning Rule, Multi-Layer Perceptrons, Back Propagation Learning Algorithm, Mathematical Model of MLP Network

 Function Approximation: Basis Function Network, Radial Basis Function Networks (RBF), MLP vs. RBF Networks, Support Vector Machine (SVM).

 Competitive Network and Associative Memory Network: Adaptive Resonance Theory (ART), ART-1 Architecture and Algorithm, Kohonen Self-Organizing Maps (SOMs), Linear Feed-Forward  Associative  Memory  Network,  Recurrent  Associative  Memory  Network, Bidirectional Associative Memory Network (BAM), Hopfield Networks.

 Fuzzy System: Introduction to Fuzzy System, Fuzzy Relations, Fuzzy Numbers, Linguistic Description and their Analytical Form, Fuzzy Control.

 Defuzzification: Defuzzification Methods, Centroid Method, Center of Sum Method, Mean of Maxima Defuzzification, Applications, Equilibrium of Learning System, Concept of Neuro-Fuzzy and Neuro-GA Network.

 Genetic Algorithm: Basic Concepts, Offspring, Encoding, Reproduction, Crossover, Mutation Operator, Application of GA

References:

B1: Neural Computing - An Introduction - R Beale and T Jackson, Publisher: Adam Hilger, 1990 IOP Publishing Ltd.

B2: C++ in Neural Networks and Fuzzy Logic - Valluru B. Rao, Publisher: M&T Books, IDG Books Worldwide, Inc. link

B3: Genetic Algorithms in Search, Optimization and Machine Learning, David E. Goldberg, Publisher: Addison-Wesley Publishing Company, INC. link

B4: An Introduction to Support Vector Machines and Other Kernel-based Learning Methods - Nello Cristianini and John Shawe-Taylor, link

B5: NEURAL NETWORKS, FUZZY LOGIC AND GENETIC ALGORITHM: SYNTHESIS AND APPLICATIONS - S. RAJASEKARAN, G. A. VIJAYALAKSHMI PAI (part1, part2, part3, part4)

Assessment & Marks Distribution: 

Quizzes/Class Test: 20 Marks* (3 best out of 4 quizzes/class tests may be taken for awarding grade)

Homework’s/Attendance: 8 Marks*

Semester Exam: 72 Marks

* - We reserve the right to change the above grading scheme.

Class test and attendance:

1. Class Test 1: Marks(B)

2. Class Test 2: Marks(B)

3. Class Test 3 and 4: Marks and Attendance (Section B) 

4. Class Test 3 and 4: Marks and Attendance (Section C)

5. Lab Quiz marks and Attendance (Section B)

6. Lab Attendance (Section C)

Class Lecture:

1. Course Outline- Lecture01

2. Introduction to the Concept: Human Brain Mechanisms and Neural Machine Intelligence.Lecture02

3.  Pattern Recognition - Lecture - 03 , Lecture - 04

4. The Basic Neuron - Lecture -05&06

5. The Multilayer Perceptron - Lecture07&09

Course Title: Sessional Based on CSE CSE 4203

Course Code: CSE 4204

Credit: 0.75 (2.5 hours per alternative week)

Lab:

1. Implementation of Nearest Neighbor classification algorithms with and without distorted pattern.  (download module01)

2. Design and implementation of single layer perceptron learning algorithm.  (download module02)

3. Design and implementation of Multi-layer Neural Networks algorithm (i.e., Back-propagation learning neural networks algorithm). (download module03)

4.  (a) Design and implementation of Kohonen Self-organizing Neural Networks algorithm.  (b) Design and implementation of Hopfield Neural Networks algorithm.  (download module04)

5. Designing a Fuzzy Logic Controller (FLC) for controlling the environmental inputs to solve the real world problem.  (download module05)

6. Design and development of a Genetic Algorithm (GA) to search and optimize a specific problem.  (download module06)

Assessment and Marks Distribution:

Attendance: 8

Quiz : 20

Lab Performance / Report: 47

Board Viva (Compulsory): 25

Total Marks: 100

Quiz marks and attendance: 

1. Quiz and Attendance -

2. Quiz and Attendance -