© Fotolia / maciek905

Specific AI Deepenings

Applying AI methods to fresh problems often requires new methods and strategies. Spatially ordered data, for example, can usually be better processed with CNNs (Convolutional Neural Networks), time series with RNNs (Recurrent neural networks), and speech data, for which longer temporal relationships need to be accounted, via so-called transformers. Depending on the availability and quality of the data sets and annotations, unsupervised and self-supervised learning strategies as well as a wide variety of data augmentation techniques are in employ. A basic understanding of these architectures, learning strategies and other AI methods enables efficient solutions to be found quickly.


The Specific AI Deepening module takes a closer look at AI architectures of varying complexity, as well as learning strategies and other specifics of machine learning. All elements of the course are accompanied with examples and interactive Jupyter notebooks. In this way, all participants have the opportunity to directly apply what they have learned.

The course is intended for all scientists who want to gain a deeper understanding of AI methods in order to make informed decisions on the way to applying them within their discipline.

Course contents:
  • CNNs
  • RNNs
  • Dimension reduction/ autoencoder
  • Attention / Transformer
  • Graph Neural Networks
  • Generalisation vs. overfitting
  • Self-supervised learning
  •  Active learning