Introduction to artificial neural networks (NS), NS logical neurons. The digital / analog Perceptron: the concept of learning with a teacher pattern recognition. Linear NS: vector spaces, auto associative memory. Multi-layer perceptron: the method of back propagation error, training and test set, generalization, selection of model validation. Hebb learning without a teacher, feature extraction, principal component analysis. Learning the competition, self-organizing map clustering, topographic display. Hybrid NS: radial-basis-function NS algorithm for training, properties. Recurrent NS: temporal structure in data, models and algorithms for training, echo state networks, recurrent self-organizing maps. Hopfield model: deterministic and stochastic dynamics, attractors in state space, auto associative memory. Deep architecture NS.

Type of methodology: Combination of lecture and hands-on

Participants receive the certificate of attendance: Yes

Paid training activity for participants: Yes, for all

Participants prerequisite knowledge: Numerical methods (linear algebra, statistics) Domain-specific background knowledge