A Closer Look at Few-shot Classification Again

Xu Luo, Hao Wu, Ji Zhang, Lianli Gao, Jing Xu, Jingkuan Song
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:23103-23123, 2023.

Abstract

Few-shot classification consists of a training phase where a model is learned on a relatively large dataset and an adaptation phase where the learned model is adapted to previously-unseen tasks with limited labeled samples. In this paper, we empirically prove that the training algorithm and the adaptation algorithm can be completely disentangled, which allows algorithm analysis and design to be done individually for each phase. Our meta-analysis for each phase reveals several interesting insights that may help better understand key aspects of few-shot classification and connections with other fields such as visual representation learning and transfer learning. We hope the insights and research challenges revealed in this paper can inspire future work in related directions. Code and pre-trained models (in PyTorch) are available at https://github.com/Frankluox/CloserLookAgainFewShot.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-luo23e, title = {A Closer Look at Few-shot Classification Again}, author = {Luo, Xu and Wu, Hao and Zhang, Ji and Gao, Lianli and Xu, Jing and Song, Jingkuan}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {23103--23123}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/luo23e/luo23e.pdf}, url = {https://proceedings.mlr.press/v202/luo23e.html}, abstract = {Few-shot classification consists of a training phase where a model is learned on a relatively large dataset and an adaptation phase where the learned model is adapted to previously-unseen tasks with limited labeled samples. In this paper, we empirically prove that the training algorithm and the adaptation algorithm can be completely disentangled, which allows algorithm analysis and design to be done individually for each phase. Our meta-analysis for each phase reveals several interesting insights that may help better understand key aspects of few-shot classification and connections with other fields such as visual representation learning and transfer learning. We hope the insights and research challenges revealed in this paper can inspire future work in related directions. Code and pre-trained models (in PyTorch) are available at https://github.com/Frankluox/CloserLookAgainFewShot.} }
Endnote
%0 Conference Paper %T A Closer Look at Few-shot Classification Again %A Xu Luo %A Hao Wu %A Ji Zhang %A Lianli Gao %A Jing Xu %A Jingkuan Song %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-luo23e %I PMLR %P 23103--23123 %U https://proceedings.mlr.press/v202/luo23e.html %V 202 %X Few-shot classification consists of a training phase where a model is learned on a relatively large dataset and an adaptation phase where the learned model is adapted to previously-unseen tasks with limited labeled samples. In this paper, we empirically prove that the training algorithm and the adaptation algorithm can be completely disentangled, which allows algorithm analysis and design to be done individually for each phase. Our meta-analysis for each phase reveals several interesting insights that may help better understand key aspects of few-shot classification and connections with other fields such as visual representation learning and transfer learning. We hope the insights and research challenges revealed in this paper can inspire future work in related directions. Code and pre-trained models (in PyTorch) are available at https://github.com/Frankluox/CloserLookAgainFewShot.
APA
Luo, X., Wu, H., Zhang, J., Gao, L., Xu, J. & Song, J.. (2023). A Closer Look at Few-shot Classification Again. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:23103-23123 Available from https://proceedings.mlr.press/v202/luo23e.html.

Related Material