Learning From Teacher’s Failure: A Reflective Learning Paradigm for Knowledge Distillation

Abstract

Knowledge Distillation transfers knowledge learned by a teacher network to a student network. A common mode of knowledge transfer is directly using the teacher network’s experience for all samples without differentiating whether the experience of teacher is successful or not. According to common sense, experience varies with its nature. Successful experience is used for guidance, and failed experience is used for correction. Inspired by that, this paper analyzes the failure of teacher and proposes a reflective learning paradigm, which additional uses heuristic knowledge extracted from the teacher’s failure besides following the authority of teacher. Specifically, this paper defines Mutual Error Distance (MED) based on the teacher’s wrong predictions. MED measures the adequacy of the decision boundary learned by teacher, which concretizes the failure of teacher. Then, this paper proposes DCGD (divide-and-conquer grouping distillation) to critically transfer the teacher’s knowledge by grouping the target task into small-scale subtasks and designing multi-branch networks on the basis of MED. Finally, a switchable training mechanism is designed to integrate a regular student which provides an option of student network without parameter addition compared with the multi-branch student network. Extensive experiments on three image classification benchmarks (CIFAR-10, CIFAR-100 and TinyImageNet) show the effectiveness of the proposed paradigm. Especially on CIFAR-100 dataset, the average error of students using DCGD+DKD decreased by 4.28%. In addition, the experiment results show that the paradigm is also applicable to self-distillation.