Continued from A5: Dependability, Adaptability and Uncertainty Quantification for Data Analysis Workflows in Large-Scale Biomedical Image Analysis
Description
We aim to refine machine learning workflows in biomedicine by enhancing annotation efficiency. Currently, the lack of high-quality training data and time-consuming manual annotations hinder progress. To address this, we will explore self-supervised representation learning, which utilizes unlabeled data, potentially reducing the need for manual annotations. A5 will focus on creating efficient frameworks to train, explore, and benchmark such representation learning approaches.
Scientists
- Marc-Andre Schulz