Krogh Group
The Krogh Group is shared between HeaDS and Department of Computer Science. It investigates methods in representation learning and generative modelling, and their applications for medical/clinical data. Representation learning is of increasing importance in many research areas since raw data is becoming more complex and thus computationally inefficient as a feature space for many applications. Common practices in representation learning, however, often use complex and large models and treat representations as a by-product of machine learning.
The group wants to revisit representation learning and generative modelling with an increased focus on explicitly learned representations. For this purpose, the group conducts research in basic machine learning as well as the application to biological data. The group is currently focusing on a method for explicit estimation of representations in deep generative decoders and its application for modeling gene expression data and clinical data. As part of the Center for Basic Machine Learning in Life Science (MLLS), the group is in close contact and constant dialogue with other machine learning research groups in Denmark.
- Novo Nordisk Foundation Collaborative Reseach Programme supports the Center for Basic Machine Learning Research in Life Science (2020) – Ole Winther and Anders Krogh
- Novo Nordisk Foundation Research Infrastructure Programme supports the National Health Data Science Sandbox for Training and Research (2020) – Anders Krogh
- Novo Nordisk Foundation Collaborative Research Programme supports SE3D: Synthetic Health Data: Ethical Development and Deployment via Deep Learning Approaches (2023) – Jennifer Bartell and Anders Krogh with Martin Bøgsted and Jan Trzaskowski at Aalborg University