Praktisches Deep Learning

What it is about

You will gain a practical introduction to Deep Learning and implement central algorithms yourself. At the end you will know various deep learning models (e.g. deep nets, CNNs, RNNs/LSTM, autoencoders, GANs, VAEs) and be able to apply them to real datasets with PyTorch, as well as improve deep-learning pipelines.

What you will be able to do

• Familiarity with various deep-learning algorithms
• Explanation of fundamental training concepts (backpropagation, over/underfitting, regularization, early stopping, cross-validation)
• Application of PyTorch for Deep Learning on real datasets
• Development and improvement of deep-learning pipelines

What the module consists of

• PraktikumCentral teaching format: exercise sheets with detailed instructions, independent or group solutions, guided exercises
• PräsentationShort (one hour) introduction to concepts and algorithms at the beginning of the sessions
• GruppenprojektFinal phase: application and deepening of concepts in group projects

Teaching method

• Kurzvortrag/PräsentationIntroduction of concepts and algorithms
• Geführte ÜbungsblätterGuided through practical implementation tasks and addressing arising questions
• Gruppenarbeit/ProjektApplication and deepening of the learned methods in realistic tasks