Publication Type
Conference Proceeding Article
Version
publishedVersion
Publication Date
6-2022
Abstract
The inductive bias of vision transformers is more relaxed that cannot work well with insufficient data. Knowledge distillation is thus introduced to assist the training of transformers. Unlike previous works, where merely heavy convolution-based teachers are provided, in this paper, we delve into the influence of models inductive biases in knowledge distillation (e.g., convolution and involution). Our key observation is that the teacher accuracy is not the dominant reason for the student accuracy, but the teacher inductive bias is more important. We demonstrate that lightweight teachers with different architectural inductive biases can be used to co-advise the student transformer with outstanding performances. The rationale behind is that models designed with different inductive biases tend to focus on diverse patterns, and teachers with different inductive biases attain various knowledge despite being trained on the same dataset. The diverse knowledge provides a more precise and comprehensive description of the data and compounds and boosts the performance of the student during distillation. Furthermore, we propose a token inductive bias alignment to align the inductive bias of the token with its target teacher model. With only lightweight teachers provided and using this cross inductive bias distillation method, our vision transformers (termed as CiT) outperform all previous vision transformers (ViT) of the same architecture on ImageNet. Moreover, our small size model CiT-SAK further achieves 82.7% Top-1 accuracy on ImageNet without modifying the attention module of the ViT.
Keywords
Adversarial attack and defense, Distillation method, Inductive bias, Performance, Representation learning, Size models, Teacher models, Teachers'
Discipline
Databases and Information Systems
Research Areas
Information Systems and Management
Publication
Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, USA, June 19-22
First Page
16752
Last Page
16761
ISBN
9781665469463
Identifier
10.1109/CVPR52688.2022.01627
Publisher
IEEE
City or Country
New Jersey
Citation
REN, Sucheng; GAO, Zhengqi; HUA, Tiany; XUE, Zihui; TIAN, Yonglong; HE, Shengfeng; and ZHAO, Hang.
Co-advise: Cross inductive bias distillation. (2022). Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, USA, June 19-22. 16752-16761.
Available at: https://ink.library.smu.edu.sg/sis_research/8538
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.
Additional URL
https://doi.org/10.1109/CVPR52688.2022.01627