Decoupled Knowledge Distillation

🤖 Plain-English Summary

advanced distillation methods are mainly based on distilling deep features from intermediate layers, while the significance of logit distillation is greatly overlooked. This paper proves the great potential of logit distillation, and we hope it will be helpful for future research.

🔑 Key Findings

• To provide a novel viewpoint to study logit distillation, we re-formulate the classical KD loss into two parts, i.e., target class knowledge distillation (TCKD) and non-target class knowledge distillation (NCKD).
• We empirically investigate and prove the effects of the two parts: TCKD transfers knowledge concerning the “difficulty” of training samples, while NCKD is the prominent reason why logit distillation works.
• More importantly, we reveal that the classical KD loss is a coupled formulation, which (1) suppresses the effectiveness of NCKD and (2) limits the flexibility to balance these two parts.

🔗 Related Resources