Single-Teacher View Augmentation: Enhancing Knowledge Distillation with Student-Guided Perturbations

Xuyi Yu, Yaohua Liu, Chengjun Li, Qiang Tang, Shuzhe Tang, Kuizhi Mei

Knowledge distillation (KD) typically relies on the fixed perspective of a single teacher, limiting the diversity of supervisory signals. While multi-teacher distillation addresses this by aggregating knowledge from multiple models, it incurs prohibitive computational and storage costs. To balance efficiency and diversity, recent research has focused on generating virtual views from a single teacher. However, existing methods face a trade-off: random perturbation approaches offer efficiency but lack controlled diversity, while structured augmentation methods require multi-stage training and incur linear parameter growth. We observe that this trade-off stems from a common design choice: using the teacher's strong but static features to generate views. Instead, we propose Shift-Augmented Knowledge Distillation (SAKD), a simple yet effective framework that leverages the student's evolving features as a dynamic condition for perturbation generation. This shift in perspective enables single-stage training while producing adaptive, diverse views through a parameter-free cyclic shift. Extensive experiments on CIFAR-100 and ImageNet demonstrate that SAKD consistently outperforms random perturbation methods and achieves accuracy on par with two-stage approaches, while using significantly fewer parameters and eliminating pre-training requirements.