Artificial intelligence is transforming biomedicine, but the success of machine learning models depends on access to large, high-quality datasets. These are often difficult to obtain in healthcare due to privacy concerns, variability and challenges in annotation. This talk will explore how synthetic data, particularly electrocardiograms (ECGs) generated through electrophysiological simulations, can be used as a powerful tool to overcome these limitations.

We will look at how synthetic data can complement real-world datasets, improve model generalization and facilitate learning in rare or underrepresented scenarios. We will provide examples to illustrate how synthetic data is being used to train deep learning models for a variety of biomedical applications.
The talk will also highlight the potential of this approach, including its scalability and controllability, while addressing key challenges such as ensuring physiological authenticity and closing the domain gap between synthetic and clinical data. Ultimately, we will explore how synthetic data can help to bridge the gap between computational modeling and real-world clinical AI applications.

This talk is held in English and part of the HeFDI Data Week 2025.

The HeFDI Data Week 2025 offers a total of twelve inputs and workshops on data management over three days - this time with a focus on artificial intelligence and data, legal issues and disciplinary tools! The events are accessible online and free of charge. They are aimed at researchers, teaching staff, students and anyone who wants to learn more about FAIR data and code, practical tools and available services. The latest developments and associated challenges will be presented and discussed with service providers and users alike.

#Digitaltag