MetaDelta: A Meta-Learning System for Few-shot Image Classification

Yudong Chen, Chaoyu Guan, Zhikun Wei, Xin Wang, Wenwu Zhu
AAAI Workshop on Meta-Learning and MetaDL Challenge, PMLR 140:17-28, 2021.

Abstract

Meta-learning aims at learning quickly on novel tasks with limited data by transferring generic experience learned from previous tasks. Naturally, few-shot learning has been one of the most popular applications for meta-learning. However, existing meta-learning algorithms rarely consider the time and resource efficiency or the generalization capacity for unknown datasets, which limits their applicability in real-world scenarios. In this paper, we propose MetaDelta, a novel practical meta-learning system for the few-shot image classification. MetaDelta consists of two core components: i) multiple meta-learners supervised by a central controller to ensure efficiency, and ii) a meta-ensemble module in charge of integrated inference and better generalization. In particular, each meta-learner in MetaDelta is composed of a unique pre-trained encoder fine-tuned by batch training and parameter-free decoder used for prediction. MetaDelta ranks first in the final phase in the AAAI 2021 MetaDL Challenge.

Cite this Paper

BibTeX
@InProceedings{pmlr-v140-chen21a, title = {MetaDelta: A Meta-Learning System for Few-shot Image Classification}, author = {Chen, Yudong and Guan, Chaoyu and Wei, Zhikun and Wang, Xin and Zhu, Wenwu}, booktitle = {AAAI Workshop on Meta-Learning and MetaDL Challenge}, pages = {17--28}, year = {2021}, editor = {Guyon, Isabelle and van Rijn, Jan N. and Treguer, Sébastien and Vanschoren, Joaquin}, volume = {140}, series = {Proceedings of Machine Learning Research}, month = {09 Feb}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v140/chen21a/chen21a.pdf}, url = {https://proceedings.mlr.press/v140/chen21a.html}, abstract = {Meta-learning aims at learning quickly on novel tasks with limited data by transferring generic experience learned from previous tasks. Naturally, few-shot learning has been one of the most popular applications for meta-learning. However, existing meta-learning algorithms rarely consider the time and resource efficiency or the generalization capacity for unknown datasets, which limits their applicability in real-world scenarios. In this paper, we propose MetaDelta, a novel practical meta-learning system for the few-shot image classification. MetaDelta consists of two core components: i) multiple meta-learners supervised by a central controller to ensure efficiency, and ii) a meta-ensemble module in charge of integrated inference and better generalization. In particular, each meta-learner in MetaDelta is composed of a unique pre-trained encoder fine-tuned by batch training and parameter-free decoder used for prediction. MetaDelta ranks first in the final phase in the AAAI 2021 MetaDL Challenge. } }
Endnote
%0 Conference Paper %T MetaDelta: A Meta-Learning System for Few-shot Image Classification %A Yudong Chen %A Chaoyu Guan %A Zhikun Wei %A Xin Wang %A Wenwu Zhu %B AAAI Workshop on Meta-Learning and MetaDL Challenge %C Proceedings of Machine Learning Research %D 2021 %E Isabelle Guyon %E Jan N. van Rijn %E Sébastien Treguer %E Joaquin Vanschoren %F pmlr-v140-chen21a %I PMLR %P 17--28 %U https://proceedings.mlr.press/v140/chen21a.html %V 140 %X Meta-learning aims at learning quickly on novel tasks with limited data by transferring generic experience learned from previous tasks. Naturally, few-shot learning has been one of the most popular applications for meta-learning. However, existing meta-learning algorithms rarely consider the time and resource efficiency or the generalization capacity for unknown datasets, which limits their applicability in real-world scenarios. In this paper, we propose MetaDelta, a novel practical meta-learning system for the few-shot image classification. MetaDelta consists of two core components: i) multiple meta-learners supervised by a central controller to ensure efficiency, and ii) a meta-ensemble module in charge of integrated inference and better generalization. In particular, each meta-learner in MetaDelta is composed of a unique pre-trained encoder fine-tuned by batch training and parameter-free decoder used for prediction. MetaDelta ranks first in the final phase in the AAAI 2021 MetaDL Challenge.
APA
Chen, Y., Guan, C., Wei, Z., Wang, X. & Zhu, W.. (2021). MetaDelta: A Meta-Learning System for Few-shot Image Classification. AAAI Workshop on Meta-Learning and MetaDL Challenge, in Proceedings of Machine Learning Research 140:17-28 Available from https://proceedings.mlr.press/v140/chen21a.html.

Related Material