Scholar
Maurice Weiler
Google Scholar ID: uQePx6EAAAAJ
University of Amsterdam
Citations & Impact
Citations
3,697
H-index
17
i10-index
17
Publications
20
Co-authors
27
list available
Publications
Resume (English only)
Academic Achievements
- Book Publication: 'Equivariant and Coordinate Independent Convolutional Networks'; Publications: Applications in biomedical and satellite image processing, environmental, chemical and material sciences, reinforcement learning, and robotics. PyTorch library escnn is widely used across various domains.
Research Experience
- Research Focus: Designing Equivariant Convolutional Neural Networks (CNNs), which are geometry-aware neural networks constrained to commute with geometric transformations of feature vector fields. During his PhD, he developed a general representation theoretic formulation of equivariant CNNs, applicable to a wide range of spaces, symmetry groups, and group actions. Additionally, he explored the generalization of CNNs to Riemannian manifolds, leading to a gauge field theory framework.
Education
- PhD: University of Amsterdam, supervised by Max Welling; Master's: Heidelberg University, Computational and Theoretical Physics.
Background
- Research Interests: Geometric deep learning, especially incorporating geometric inductive priors into deep neural networks. Professional Field: Computational and theoretical physics.
Miscellany
- Personal Interests: Climbing, hiking, mountain biking, cooking, playing strategic games, and DJing.
