Learning Student Networks via Feature Embedding
This work addresses the challenge of reducing storage and computational demands for mobile AI applications, offering an incremental improvement over existing teacher-student methods.
The paper tackles the problem of deploying deep neural networks on mobile devices by proposing a knowledge distillation method that transfers knowledge from a teacher to a student network without adding extra parameters, using a feature embedding approach with a locality preserving loss. Experiments show it outperforms state-of-the-art methods in computational and storage complexity on benchmark datasets.
Deep convolutional neural networks have been widely used in numerous applications, but their demanding storage and computational resource requirements prevent their applications on mobile devices. Knowledge distillation aims to optimize a portable student network by taking the knowledge from a well-trained heavy teacher network. Traditional teacher-student based methods used to rely on additional fully-connected layers to bridge intermediate layers of teacher and student networks, which brings in a large number of auxiliary parameters. In contrast, this paper aims to propagate information from teacher to student without introducing new variables which need to be optimized. We regard the teacher-student paradigm from a new perspective of feature embedding. By introducing the locality preserving loss, the student network is encouraged to generate the low-dimensional features which could inherit intrinsic properties of their corresponding high-dimensional features from teacher network. The resulting portable network thus can naturally maintain the performance as that of the teacher network. Theoretical analysis is provided to justify the lower computation complexity of the proposed method. Experiments on benchmark datasets and well-trained networks suggest that the proposed algorithm is superior to state-of-the-art teacher-student learning methods in terms of computational and storage complexity.