Multi-Mode Online Knowledge Distillation for Self-Supervised Visual Representation Learning
This work addresses the challenge of improving representation learning for small models in self-supervised learning, though it appears incremental as it builds on existing SSL-KD methods by introducing collaborative learning.
The paper tackles the problem of boosting self-supervised visual representation learning by proposing a multi-mode online knowledge distillation method (MOKD), where two models learn collaboratively through self-distillation and cross-distillation modes, resulting in both models outperforming their independently trained baselines and existing SSL-KD methods.
Self-supervised learning (SSL) has made remarkable progress in visual representation learning. Some studies combine SSL with knowledge distillation (SSL-KD) to boost the representation learning performance of small models. In this study, we propose a Multi-mode Online Knowledge Distillation method (MOKD) to boost self-supervised visual representation learning. Different from existing SSL-KD methods that transfer knowledge from a static pre-trained teacher to a student, in MOKD, two different models learn collaboratively in a self-supervised manner. Specifically, MOKD consists of two distillation modes: self-distillation and cross-distillation modes. Among them, self-distillation performs self-supervised learning for each model independently, while cross-distillation realizes knowledge interaction between different models. In cross-distillation, a cross-attention feature search strategy is proposed to enhance the semantic feature alignment between different models. As a result, the two models can absorb knowledge from each other to boost their representation learning performance. Extensive experimental results on different backbones and datasets demonstrate that two heterogeneous models can benefit from MOKD and outperform their independently trained baseline. In addition, MOKD also outperforms existing SSL-KD methods for both the student and teacher models.