TY - RPRT TI - Regularising for invariance to data augmentation improves supervised learning AU - Aleksander Botev AU - Matthias Bauer AU - Soham De PY - 2022 UR - https://arxiv.org/abs/2203.03304 ID - 2203.03304 ER -