Knowledge Distillation (KD) is a model-agnostic technique to improve model quality while having a fixed capacity budget. It is a commonly used technique for model compression, where a larger capacity teacher model with better quality is used to train a more compact student model with better inference efficiency. Through distillation, one hopes to benefit from students compactness, without sacrificing too much on model quality. Despite the large success of knowledge distillation, better understanding of how it benefits student models training dynamics remains under-explored. In this paper, we categorize teachers knowledge into three hierarchical levels and study its effects on knowledge distillation: (1) knowledge of the `universe, where KD brings a regularization effect through label smoothing; (2) domain knowledge, where teacher injects class relationships prior to students logit layer geometry; and (3) instance specific knowledge, where teacher rescales student models per-instance gradients based on its measurement on the event difficulty. Using systematic analyses and extensive empirical studies on both synthetic and real-world datasets, we confirm that the aforementioned three factors play a major role in knowledge distillation. Furthermore, based on our findings, we diagnose some of the failure cases of applying KD from recent studies.
تحميل البحث