SYNTHIA is presenting at the 39th IEEE International Symposium on Computer-Based Medical Systems (CBMS 2026), 3-5 June 2026, in Limassol, Cyprus.
SYNTHIA partner Matteo Bonazzi, UNIBO, is presenting the abstract "Benchmarking Tabular Generative Models for Clinical Data Synthesis" in Session 2E: Special Track on Synthetic Healthcare Data Generation and Clinical Decision Support 1: Federated Learning & Structured Data Generation.
This publication evaluates several state-of-the-art tabular generative models on a real-world clinical dataset from the Alzheimer's Disease Neuroimaging Initiative (ADNI). The study compares GAN-, VAE-, diffusion-, and tree-based approaches across key criteria such as data fidelity, statistical similarity, privacy, and computational efficiency.
Results show that no single model dominates across all metrics. Forest Diffusion achieved strong distributional fidelity, while TVAE offered a good balance between performance and training efficiency. Overall, the findings highlight the importance of carefully selecting synthetic data generation methods based on specific clinical requirements.
About IEEE CBMS 2026
The 39th IEEE International Symposium on Computer-Based Medical Systems (CBMS 2026) will take place in Limassol, Cyprus, bringing together clinicians, researchers, engineers, and developers from around the world to share the latest advances at the intersection of medicine and computing.
Recognized as a leading international forum in the field, CBMS provides a multidisciplinary platform for presenting innovative research, discussing emerging technologies, and addressing current challenges in computer-based medical systems. The 2026 edition will be hosted in Limassol, a vibrant coastal city offering modern conference facilities alongside Cyprus' rich history, culture, and hospitality.
