EUVIMEDEUVIMED entdecken
Crossref

Online Knowledge Distillation via Mutual Contrastive Learning for Visual Recognition

Chuanguang Yang, Zhulin An, Helong Zhou, Fuzhen Zhuang, Yongjun Xu, Qian Zhang

IEEE Transactions on Pattern Analysis and Machine Intelligence · 2023-08 · 45 (8) · 10212-10227

DOI: 10.1109/tpami.2023.3257878 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023-08. Verfasst von Chuanguang Yang und 5 weiteren Autor:innen. Die Arbeit wurde bisher 108-mal zitiert (Crossref) und nennt 60 Literaturangaben. Gefördert von: National Key Research and Development Program of China, National Natural Science Foundation of China, Fundamental Research Funds for the Central Universities.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The teacher-free online Knowledge Distillation (KD) aims to train an ensemble of multiple student models collaboratively and distill knowledge from each other. Although existing online KD methods achieve desirable performance, they often focus on class probabilities as the core knowledge type, ignoring the valuable feature representational information. We present a Mutual Contrastive Learning (MCL) framework for online KD. The core idea of MCL is to perform mutual interaction and transfer of contrastive distributions among a cohort of networks in an online manner. Our MCL can aggregate cross-network embedding information and maximize the lower bound to the mutual information between two networks. This enables each network to learn extra contrastive knowledge from others, leading to better feature representations, thus improving the performance of visual recognition tasks. Beyond the final layer, we extend MCL to intermediate layers and perform an adaptive layer-matching mechanism trained by meta-optimization. Experiments on image classification and transfer learning to visual recognition tasks show that layer-wise MCL can lead to consistent performance gains against state-of-the-art online KD approaches. The superiority demonstrates that layer-wise MCL can guide the network to generate better feature representations. Our code is publicly avaliable at https://github.com/winycg/L-MCL.

Förderung

  • National Key Research and Development Program of China
  • National Natural Science Foundation of China
  • Fundamental Research Funds for the Central Universities

Zitieren

APA

Yang, C., An, Z., Zhou, H., Zhuang, F., Xu, Y., & Zhang, Q. (2023). Online Knowledge Distillation via Mutual Contrastive Learning for Visual Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(8), 10212-10227. https://doi.org/10.1109/tpami.2023.3257878

Vancouver

Yang C, An Z, Zhou H, Zhuang F, Xu Y, Zhang Q. Online Knowledge Distillation via Mutual Contrastive Learning for Visual Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2023;45(8):10212-10227. doi:10.1109/tpami.2023.3257878

BibTeX
@article{Yang2023,
  author = {Yang, Chuanguang and An, Zhulin and Zhou, Helong and Zhuang, Fuzhen and Xu, Yongjun and Zhang, Qian},
  title = {Online Knowledge Distillation via Mutual Contrastive Learning for Visual Recognition},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2023},
  volume = {45},
  number = {8},
  pages = {10212--10227},
  doi = {10.1109/tpami.2023.3257878}
}
Kostenlos starten →