Deep Learning

Course description:
The course presents common and state-of-the-art Deep Neural Network architectures, including: Multilayer Perceptrons, Convolutional architectures, Recurrent Neural Networks, Attention mechanisms and Transformers. In addition, the course discusses training related concepts and issues that are crucial for gaining a firm understanding of the underlying principles of the most important training procedures, like the mechanics of the backpropagation algorithm, parameter tuning and convergence/overfitting issues. The course evolves across five lectures and student progress is assessed by means of a team project deliverable.

Course coordinator: Prof. Aggelos Pikrakis