Upside down, a cat's still a cat: Evolving image recognition with Geometric Deep Learning

In this first in a series of posts on group-equivariant convolutional neural networks (GCNNs), meet the main actors — groups — and concepts (equivariance).