Profil
The Machine Learning and Uncertainty group 8.44 develops and validates machine learning techniques with applications in neuroscience and medicine. Current research interests include
- Solutions to the brain's electromagnetic inverse problem using Bayesian methods
- Brain functional connectivity analysis using electro- and magnetoencephalography
- Interpretation/explanation of machine learning models
- Creation of synthetic reference datasets and benchmarking procedures to assess quality aspects of computational methods
- Uncertainty estimates for inverse problems
Current applications include
- Diagnosis and characterization of psychiatric and neurological disorders
- Intensive care unit (ICU) data
- Brain-computer interfaces
The group is tightly linked with the Brain and Data Science Group at Charité - Universitätsmedizin Berlin and the Department of Uncertainty, Inverse Modeling and Machine Learning at Technische Universität Berlin, both also led by Dr. Stefan Haufe.
Forschung/Entwicklung
Currently, the following projects are carried out at PTB
Towards standardized quality control for AI systems in critical care
Machine learning and uncertainty quantification for bioelectromagnetic inverse solutions and signal separation methods
Advancing the theory and practice of machine learning model explanations in biomedicine
Development and quality assessment of neuroimaging-based biomarkers for brain disorders
Further projects, carried out at TUB and/or Charité, are
Dienstleistungen
The group currently does not offer any specific service. However, we are open to discussing machine learning and data science problems arising in other PTB groups.