Deep Learning

This group focuses on methodological and applied research in the broader context of Deep Learning (DL). Current projects include

• Semi-structured Distributional Deep Learning (SDDL), a novel research field focusing on unifying regression models and deep neural networks to learn entire distributions. The approach estimates the statistical model part within a neural network while ensuring interpretability and thereby also allows estimation of classical regression models in high-dimensional settings.

• Semi-supervised Deep Learning for Time-series

• Incorporation of uncertainty in the design of personalized vaccines for cancer (in collaboration with the Institute of Computational Biology at the Helmholtz Zentrum)

• Deep representation learning from imbalanced data

• Unsupervised Meta-Learning for Imbalanced and Out-of-distribution Tasks

• Gaussian Processes, Uncertainty quantification and Variational inference

Members