Hierarchical Multi-Attention Transfer for Knowledge Distillation

Jianping Gou, Liyuan Sun, Baosheng Yu, Shaohua Wan, Dacheng Tao

Abstract: Knowledge distillation (KD) is a powerful and widely applicable technique for the compression of deep learning models. The main idea of knowledge distillation is to transfer knowledge from a large teacher model to a small student model, where the attention mechanism has been intensively explored in regard to its great flexibility for managing different teacher-student architectures. However, existing attention-based methods usually transfer similar attention knowledge from the intermediate layers of deep neural networks, leaving the hierarchical structure of deep representation learning poorly investigated for knowledge distillation. In this paper, we propose a hierarchical multi-attention transfer framework (HMAT), where different types of attention are utilized to transfer the knowledge at different levels of deep representation learning for knowledge distillation. Specifically, position-based and channel-based attention knowledge characterize the knowledge from low-level and high-level feature representations, respectively, and activation-based attention knowledge characterize the knowledge from both mid-level and high-level feature representations. Extensive experiments on three popular visual recognition tasks, image classification, image retrieval, and object detection, demonstrate that the proposed hierarchical multi-attention transfer or HMAT significantly outperforms recent state-of-the-art KD methods.