Channel Importance Matters in Few-Shot Image Classification

Xu Luo, Jing Xu, Zenglin Xu
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:14542-14559, 2022.

Abstract

Few-Shot Learning (FSL) requires vision models to quickly adapt to brand-new classification tasks with a shift in task distribution. Understanding the difficulties posed by this task distribution shift is central to FSL. In this paper, we show that a simple channel-wise feature transformation may be the key to unraveling this secret from a channel perspective. When facing novel few-shot tasks in the test-time datasets, this transformation can greatly improve the generalization ability of learned image representations, while being agnostic to the choice of datasets and training algorithms. Through an in-depth analysis of this transformation, we find that the difficulty of representation transfer in FSL stems from the severe channel bias problem of image representations: channels may have different importance in different tasks, while convolutional neural networks are likely to be insensitive, or respond incorrectly to such a shift. This points out a core problem of the generalization ability of modern vision systems which needs further attention in the future.

Cite this Paper


BibTeX
@InProceedings{pmlr-v162-luo22c, title = {Channel Importance Matters in Few-Shot Image Classification}, author = {Luo, Xu and Xu, Jing and Xu, Zenglin}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {14542--14559}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/luo22c/luo22c.pdf}, url = {https://proceedings.mlr.press/v162/luo22c.html}, abstract = {Few-Shot Learning (FSL) requires vision models to quickly adapt to brand-new classification tasks with a shift in task distribution. Understanding the difficulties posed by this task distribution shift is central to FSL. In this paper, we show that a simple channel-wise feature transformation may be the key to unraveling this secret from a channel perspective. When facing novel few-shot tasks in the test-time datasets, this transformation can greatly improve the generalization ability of learned image representations, while being agnostic to the choice of datasets and training algorithms. Through an in-depth analysis of this transformation, we find that the difficulty of representation transfer in FSL stems from the severe channel bias problem of image representations: channels may have different importance in different tasks, while convolutional neural networks are likely to be insensitive, or respond incorrectly to such a shift. This points out a core problem of the generalization ability of modern vision systems which needs further attention in the future.} }
Endnote
%0 Conference Paper %T Channel Importance Matters in Few-Shot Image Classification %A Xu Luo %A Jing Xu %A Zenglin Xu %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-luo22c %I PMLR %P 14542--14559 %U https://proceedings.mlr.press/v162/luo22c.html %V 162 %X Few-Shot Learning (FSL) requires vision models to quickly adapt to brand-new classification tasks with a shift in task distribution. Understanding the difficulties posed by this task distribution shift is central to FSL. In this paper, we show that a simple channel-wise feature transformation may be the key to unraveling this secret from a channel perspective. When facing novel few-shot tasks in the test-time datasets, this transformation can greatly improve the generalization ability of learned image representations, while being agnostic to the choice of datasets and training algorithms. Through an in-depth analysis of this transformation, we find that the difficulty of representation transfer in FSL stems from the severe channel bias problem of image representations: channels may have different importance in different tasks, while convolutional neural networks are likely to be insensitive, or respond incorrectly to such a shift. This points out a core problem of the generalization ability of modern vision systems which needs further attention in the future.
APA
Luo, X., Xu, J. & Xu, Z.. (2022). Channel Importance Matters in Few-Shot Image Classification. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:14542-14559 Available from https://proceedings.mlr.press/v162/luo22c.html.

Related Material