Distilling Knowledge via Intermediate Classifiers

TL;DR

This paper proposes a novel knowledge distillation method using intermediate classifiers at various depths of the teacher model to better transfer knowledge to resource-limited students, especially when there is a large capacity gap.

Contribution

Introducing intermediate classifier heads at different depths of the teacher to create a cohort of heterogeneous teachers for improved knowledge transfer.

Findings

Outperforms canonical knowledge distillation methods.

Effective across various teacher-student pairs and datasets.

Mitigates capacity gap issues in knowledge transfer.

Abstract

The crux of knowledge distillation is to effectively train a resource-limited student model with the guide of a pre-trained larger teacher model. However, when there is a large difference between the model complexities of teacher and student (i.e., capacity gap), knowledge distillation loses its strength in transferring knowledge from the teacher to the student, thus training a weaker student. To mitigate the impact of the capacity gap, we introduce knowledge distillation via intermediate heads. By extending the intermediate layers of the teacher (at various depths) with classifier heads, we cheaply acquire a cohort of heterogeneous pre-trained teachers. The intermediate classifier heads can all together be efficiently learned while freezing the backbone of the pre-trained teacher. The cohort of teachers (including the original teacher) co-teach the student simultaneously. Our experiments on various teacher-student pairs and datasets have demonstrated that the proposed approach outperforms the canonical knowledge distillation approach and its extensions.