Towards Learning Group-Equivariant Features for Domain Adaptive 3D Detection | Read Paper on Bytez