Neural Networks and Applications (EE319)
Introduction: Structure of the human brain, Organization of the brain, Biological neuron, Mc-Culloch-Pitts neuron model. Various thresholding functions, Feature vectors and feature space. Classification techniques – nearest neighbour classification. Distance metrics, Linear classifiers, Decision regions. The single layer and multilayer perception, Multilayer perception algorithm, Solution of the XOR problem, Visualizing the network behaviour in terms of energy functions, Mexican Hat function. Learning in Neural networks, supervised and unsupervised learning, Feed-forward networks, Linearly non-separable pattern classification, Delta learning rule. Error back-propagation training algorithms, Feedback networks – Hopfield network, The energy landscape, Storing patterns, Recall phase, The Boltzmann machine, The traveling salesman problem. Associative memories – basic concepts, Recurrent autoassociative memory, Retrieval and Storage algorithm, Stability considerations. Application of neural systems – Linear programming, Modeling networks, Character recognition, Control system applications, Robotic applications.