RotoGrad - Gradient Homogenization in Multitask Learning
Gradient Equivalence in Siamese Self-Supervised Learning