Syllabus - Neural Networks & Fuzzy Logic (EI 703 (C))

ARCHITECTURE OF NEURAL NETWORKS

Architectures: motivation for the development of natural networks-artificial neural networks-biological neural networks-area of applications-typical Architecture-setting weights-common activations functionsBasic learning rules- Mcculloch-Pitts neuron- Architecture, algorithm, applications-single layer net for Page 1 of 7 pattern classification- Biases and thresholds, linear separability - Hebb’srule- algorithm -perceptron - Convergence theorem-Delta rule

BASIC NEURAL NETWORK TECHNIQUES

Back propagation neural net:standard back propagation-architecture algorithm- derivation of learning rulesnumber of hidden layers--associative and other neural networks- hetro associative memory neural net, auto associative net- Bidirectional associative memory-applications-Hopfield nets-Boltzman machine

COMPETITIVE NEURAL NETWORKS

Neural network based on competition: fixed weight competitive nets- Kohonenself organizing maps and applications-learning vector quantization-counter propagation nets and applications adaptive resonance theory: basic architecture and operation-architecture, algorithm, application and analysis of ART1 & ART2

SPECIAL NEURAL NETWORKS

Cognitron and Neocognitron - Architecture, training algorithm and application-fuzzy associate memories, fuzzy system architecture- comparison of fuzzy and neural systems.

FUNDAMENTALS OF FUZZY LOGIC

Basic concepts: fuzzy set theory- basic concept of crisp sets and fuzzy sets- complements- unionintersection- combination of operation- general aggregation operations- fuzzy relations- compatibility relations-orderings- morphisms- fuzzy relational equations-fuzzy set and systems