Artificial Neural Networks and Machine Learning
|Class Hours - Lab Hours||3 - 1|
|Lecturers||Stefanos Kollias, Andreas-Georgios Stafylopatis, Giorgos Stamou, Georgios Alexandridis (T & R Associates), Georgios Siolas (T & R Associates), Paraskevi Tzouveli (T & R Associates)|
The course covers topics from the area of neural networks and other techniques from the broader area of computational intelligence, such as fuzzy systems, genetic algorithms and hybrid approaches: Neural network models and architectures, learning procedures, dynamic behavior, convergence and stability. Feedforward networks and learning through error correction (multi-layer perceptron, back-propagation algorithm), associative networks (Hopfield, BAM), recurrent multi-layer networks, competitive learning networks (Kohonen maps, ART models), local learning rules (RBF networks) support vector machines, combinations of neural networks (ensembles). Applications (pattern recognition, signal/image processing, control and robotics, diagnosis, prediction, optimization). Implementations (parallelism, VLSI). Hybrid systems (fuzzy neural systems, evolutionary neural networks).