Knowledge distillation ; NLP: pretraining AND ML/NLP blog
Common descendants