Advances in MetaDL: AAAI 2021 Challenge and Workshop

Adrian El Baz, Isabelle Guyon, Zhengying Liu, Jan N. van Rijn, Sebastien Treguer, Joaquin Vanschoren
AAAI Workshop on Meta-Learning and MetaDL Challenge, PMLR 140:1-16, 2021.

Abstract

To stimulate advances in meta-learning using deep learning techniques (MetaDL), we organized in 2021 a challenge and an associated workshop. This paper presents the design of the challenge and its results, and summarizes presentations made at the workshop. The challenge focused on few-shot learning classification tasks of small images. Participants’ code submissions were run in a uniform manner, under tight computational constraints. This put pressure on solution designs to use existing architecture backbones and/or pre-trained networks. Winning methods featured various classifiers trained on top of the second last layer of popular CNN backbones, fined-tuned on the meta-training data (not necessarily in an episodic manner), then trained on the labeled support and tested on the unlabeled query sets of the meta-test data.

Cite this Paper

BibTeX
@InProceedings{pmlr-v140-el-baz21a, title = {Advances in MetaDL: AAAI 2021 Challenge and Workshop}, author = {El Baz, Adrian and Guyon, Isabelle and Liu, Zhengying and van Rijn, Jan N. and Treguer, Sebastien and Vanschoren, Joaquin}, booktitle = {AAAI Workshop on Meta-Learning and MetaDL Challenge}, pages = {1--16}, 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/el-baz21a/el-baz21a.pdf}, url = {https://proceedings.mlr.press/v140/el-baz21a.html}, abstract = {To stimulate advances in meta-learning using deep learning techniques (MetaDL), we organized in 2021 a challenge and an associated workshop. This paper presents the design of the challenge and its results, and summarizes presentations made at the workshop. The challenge focused on few-shot learning classification tasks of small images. Participants’ code submissions were run in a uniform manner, under tight computational constraints. This put pressure on solution designs to use existing architecture backbones and/or pre-trained networks. Winning methods featured various classifiers trained on top of the second last layer of popular CNN backbones, fined-tuned on the meta-training data (not necessarily in an episodic manner), then trained on the labeled support and tested on the unlabeled query sets of the meta-test data.} }
Endnote
%0 Conference Paper %T Advances in MetaDL: AAAI 2021 Challenge and Workshop %A Adrian El Baz %A Isabelle Guyon %A Zhengying Liu %A Jan N. van Rijn %A Sebastien Treguer %A Joaquin Vanschoren %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-el-baz21a %I PMLR %P 1--16 %U https://proceedings.mlr.press/v140/el-baz21a.html %V 140 %X To stimulate advances in meta-learning using deep learning techniques (MetaDL), we organized in 2021 a challenge and an associated workshop. This paper presents the design of the challenge and its results, and summarizes presentations made at the workshop. The challenge focused on few-shot learning classification tasks of small images. Participants’ code submissions were run in a uniform manner, under tight computational constraints. This put pressure on solution designs to use existing architecture backbones and/or pre-trained networks. Winning methods featured various classifiers trained on top of the second last layer of popular CNN backbones, fined-tuned on the meta-training data (not necessarily in an episodic manner), then trained on the labeled support and tested on the unlabeled query sets of the meta-test data.
APA
El Baz, A., Guyon, I., Liu, Z., van Rijn, J.N., Treguer, S. & Vanschoren, J.. (2021). Advances in MetaDL: AAAI 2021 Challenge and Workshop. AAAI Workshop on Meta-Learning and MetaDL Challenge, in Proceedings of Machine Learning Research 140:1-16 Available from https://proceedings.mlr.press/v140/el-baz21a.html.

Related Material