NEURAL NETWORKS (18EC752)

NEURAL NETWORKS

Course Code : 18EC752
CIE Marks :40
Lecture Hours/Week : 03
SEE Marks :60
Total Number of Lecture Hours 2 40 (08 Hrs / Module)
Exam Hours :03
CREDITS — 03

Course Learning Objectives: This course will enable students to:

• Understand the basics of ANN and comparison with Human brain.
• Acquire knowledge on Generalization and function approximation of various ANN architectures.
• Understand reinforcement learning using neural networks
• Acquire knowledge of unsupervised learning using neural networks.

Module -1

Introduction: Biological Neuron —Artificial Neural Model -Types of activation functions — Architecture: Feedforward and Feedback, Convex Sets, Convex Hull and Linear Separability, Non-Linear Separable Problem. XOR Problem, Multilayer Networks. Learning: Learning Algorithms, Error correction and Gradient Descent Rules, Learning objective of TLNs, Perceptron Learning Algorithm, Perceptron Convergence Theorem. L1,L2

Module -2

Supervised Learning: Perceptron learning and Non Separable sets, (x-Least Mean Square Learning, MSE Error surface, Steepest Descent Search, u-LMS approximate to gradient descent, Application of LMS to Noise Cancelling, Multi-layered Network Architecture, Backpropagation Learning Algorithm, Practical consideration of BP algorithm. L1,L2,L3

Module -3

Support Vector Machines and Radial Basis Function: Learning from Examples, Statistical Learning Theory, Support Vector Machines, SVM application to Image Classification, Radial Basis Function Regularization theory, Generalized RBF Networks, Learning in RBFNs, RBF application to face recognition.

Module -4

Attractor Neural Networks: Associative Learning Attractor Associative Memory, Linear Associative memory, Hopfield Network, application of Hopfield Network, BrainState inaBox neuralNetworkflimulated AnnealingBoltzmanL Machine, Bidirectional Associative Memory. L1,L2,L3

Module -5

Self -organization Feature Map: Maximal Eigenvector Filtering, Extracting Principal Components, Generalized Learning Laws, Vector Quantization, Self organization Feature Maps, Application of SOM, Growing Neural Gas. L1,L2,L3

Course Outcomes: At the end of the course, students should be able to:

1. Describe the basics of ANN and comparison with Human brain.
2. Understand the role of neural networks in engineering, artificial intelligence, and cognitive modelling.
3. Understand the concepts and techniques of neural networks through the study of the most important neural network models.
4. Evaluate whether neural networks are appropriate to a particular application.
5. Apply neural networks to particular application, and to know what steps to take to improve performance.

Question paper pattern:

• The examination will be conducted for 100 marks with a question paper containing 10 full questions, each of 20 marks.
• Each filll question can have a maximum of 4 sub-questions.
• There will be 2 full questions from each module covering all the topics of the module. 
• Students will have to answer 5 full questions, selecting one full question from each module.
• The total marks will be proportionally reduced to 60 marks as SEE marks is 60.

Text Book:

• Neural Networks A Classroom Approach —Satish Kumar, McGraw Hill Education (India) Pvt. Ltd, Second Edition.

Reference Books:

1. Introduction to Artificial Neural Systems - J .M. Zurada, Jaico Publications 1994.
2. Artificial Neural Networks— B. Yegnanarayana, PHI, New Delhi 1998.