Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges

Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges provides a unified framework for understanding and developing deep learning architectures based on geometric principles such as symmetry and invariance.

It covers key concepts like grids, groups, graphs, geodesics, and gauges, offering insights into how these principles underlie successful architectures like CNNs and graph neural networks.