DAWs for Annotation Efficient Machine Learning in Biomedical Imaging Research (A05)

• Subject Area: Cognitive, Systems and Behavioural Neurobiology ; Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing ; Data Management, Data-Intensive Systems, Computer Science Methods in Business Informatics
• Term: since 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 414984028

The goal is to enhance infrastructures for machine learning (ML)-heavy DAWs in pioneering biomedical imaging applications. In Phase I, we focused on building a framework for adaptable image analysis workflows based on transfer learning and uncertainty quantification. However, the lack of high-quality annotated data remains severe. In Phase II, we therefore will focus on a scalable DAW-integration of recent self-supervised representation learning techniques. We will develop methods for determining the optimal combination of imaging modalities and representation learning techniques, estimating the number of required annotations, and integrating manual annotations during training of ML-heavy DAWs.

• DFG Programme: Collaborative Research Centres
• Subproject of: SFB 1404: FONDA - Foundations of Workflows for Large-Scale Scientific Data Analysis
• Applicant Institution: Humboldt-Universität zu Berlin
• Project Heads: Professorin Dr. Dagmar Kainmüller; Professorin Dr. Kerstin Ritter