Artificial Neural Networks

Course Objectives :                           Why this Subject?

1. Impart the knowledge of neural network, models of neuron and knowledge representation
2. Analyse single layer and multi layer perceptrons and also describe application areas of each of them.
3. Understand the importance of back propagation and self organization maps.
4. Describe neuro dynamics and Hopfiled models.

Course Outcomes (CO)

• CO1 Compare the similarity between of Biological networks and Neural networks
• CO2 Perform the training of neural networks using various learning rules.
• CO3 Describe the concepts of forward and backward propagations
• CO4 Analyze and construct Hopfield Models

UNIT-I

Introduction: A Neural Network, Human Brain, Models of a Neuron, Neural Networks viewed as Directed Graphs, Network Architectures, Knowledge Representation, Artificial Intelligence and Neural Networks

Learning Process: Error Correction Learning, Memory Based Learning, Hebbian Learning, Competitive, Boltzmann Learning, Credit Assignment Problem, Memory, Adaption, Statistical Nature of the Learning Process

UNIT-II

Single Layer Perceptrons: Adaptive Filtering Problem, Unconstrained Organization Techniques, Linear Least Square Filters, Least Mean Square Algorithm, Learning Curves, Learning Rate Annealing Techniques, Perceptron –Convergence Theorem, Relation Between Perceptron and Bayes Classifier for a Gaussian Environment

Multilayer Perceptron: Back Propagation Algorithm XOR Problem, Heuristics, Output Representation and Decision Rule, Computer Experiment, Feature Detection

UNIT-III

Back Propagation: Back Propagation and Differentiation, Hessian Matrix, Generalization, Cross Validation, Network Pruning Techniques, Virtues and Limitations of Back Propagation Learning, Accelerated Convergence, Supervised Learning

UNIT – IV

Self-Organization Maps (SOM): Two Basic Feature Mapping Models, Self-Organization Map, SOM Algorithm, Properties of Feature Map, Computer Simulations, Learning Vector Quantization, Adaptive Patter Classification

Neuro Dynamics: Dynamical Systems, Stability of Equilibrium States, Attractors, Neuro Dynamical Models, Manipulation of Attractors as a Recurrent Network Paradigm Hopfield Models – Hopfield Models, Computer Experiment

Textbook(s):

1. Introduction to Artificial Neural Systems by Jacek M.Zurada. Jaico Publ.House.
2. Neural networks in Computer intelligence, Li Min Fu TMH 2003.

References:

1. “Neural Networks :A Comprehensive formulation”, Simon Haykin, 1998, AW
2. “Neural Network Fundamentals” – N.K. Bose , P. Liang, 2002, T.M.H