Supervision: Michaël Defferrard
Project type:
Semester project (master)
Finished
The goal of the project is to empirically study the difference in performance of Neural Networks that are either equivariant or invariant to rotation, both on the plane and the sphere.