Max Welling
Papers
3
Total Citations
62
H-Index
3
About
Max Welling is a leading figure in machine learning, renowned for his pioneering work in geometric deep learning and probabilistic modeling. His research focuses on developing algorithms that respect the underlying symmetries and structures of data, particularly for non-Euclidean domains like spheres and manifolds. A key contribution is the introduction of Spherical CNNs (2018, 38 citations), which extended convolutional neural networks to spherical images, enabling breakthroughs in omnidirectional vision for drones and robotics. Welling also advanced 3D point cloud processing with SVNet (2022, 12 citations), which combines SO(3) equivariance with model binarization for efficient, robust performance on edge devices. In probabilistic modeling, his work on Harmonic Exponential Families on Manifolds (2015, 12 citations) provides flexible, fast-to-train distributions for data on curved spaces, addressing challenges in geosciences and robotics. A former Distinguished Scientist at Google and current professor at the University of Amsterdam, Welling’s contributions have shaped modern AI, earning him a reputation as a visionary in learning on non-Euclidean data.
Research Focus
Key Achievements
Top Papers
- 1Spherical CNNs38 citations · 2018
- 2
- 3Harmonic Exponential Families on Manifolds12 citations · 2015