| Kommentar |
Deep learning has become an omnipresent topic in multiple fields of research. This seminar is designed to provide a general introduction into the theoretical and practical principles of deep learnig, i.e. deep neural networks, covering computation graphs, loss functions, regularization and data augmentation amongst others. Starting with basic network architectures, different modern deep neural networks i.e. graph neural networks, and deep generative models, i.e. auto-encoders, will be covered.
Organizational matters: The seminar is restricted to a maximum of 20 participants. Part of the seminar is an ungraded presentation. The course Multivariate Statistical Analysis, Econometric Methods or comparable, good good statistical programming skills are required as a prerequisite.
Registration matters and general introduction into the available topics will be present in the first meeting. If there are more registrations than places, the decision will be made according to the rules of the HU ZSP (lottery). |
| Bemerkung |
StO/PO MA 2016: 6 LP, Modul: "Research Seminar in Data Science"
StO/PO MEMS 2016: 6 LP, Modul: "Research Seminar in Data Science", Major: Quantitative Methods |
