Learning SO(3) Equivariant Representations with Spherical CNNs

Mathematics

• 3D Model 50%
• Benchmark 35%
• Convolution 29%
• Data Augmentation 95%
• Equivariance 50%
• Equivariant 100%
• Filter 31%
• Learning 97%
• Multivalued Functions 47%
• Neural Networks 33%
• Performance 21%
• Pooling 47%
• Retrieval 45%
• Spectrality 26%
• Spherical Functions 43%
• Spherical Harmonics 42%
• Standards 19%
• Symmetry 24%

Engineering & Materials Science

• Convolution 52%
• Convolutional neural networks 48%