Bas Veeling

Timeline

1. 2018

   PatchCamelyon for pathology

   In 2018, Veeling and his collaborators publish PatchCamelyon (PCam), a dataset that turns metastasis detection into a straightforward binary image-classification task for machine-learning research.

2. 2019

   A NeurIPS honorable mention

   In 2019, he coauthors “Putting An End to End-to-End: Gradient-Isolated Learning of Representations” with Sindy Löwe and Peter O’Connor. The paper proposes local self-supervised learning through gradient-isolated modules and earns a NeurIPS Outstanding New Directions Paper honorable mention.

3. 2019

   Uncertainty for medical segmentation

   In 2019, he coauthors “Supervised Uncertainty Quantification for Segmentation with Multiple Annotations,” which treats variation among graders as a supervised signal for uncertainty estimates in medical image segmentation.

4. 2023

   PDE-Refiner for neural solvers

   In 2023, he coauthors PDE-Refiner for neural partial differential equation solvers. The diffusion-inspired model uses multistep refinement to model all frequency components more accurately.

5. July 2024

   A PhD in medical deep learning

   In July 2024, he completes a PhD at the University of Amsterdam’s Amsterdam Machine Learning Lab under Max Welling and M. Petković. His thesis applies deep learning to medical image analysis and addresses expert knowledge, unlabeled data and predictive uncertainty.

6. February 2025

   BioEmu-1 for protein ensembles

   In February 2025, he coauthors Microsoft Research’s introduction of BioEmu-1, a deep-learning model that generates thousands of protein structures per hour on a single GPU.

7. As of October 2026

   Senior research engineer at Microsoft

   As of October 2026, he is a Senior Research Engineer at Microsoft Research Amsterdam, where he works on machine learning for molecular simulation.