Transfer learning based few-shot classification using optimal transport mapping from preprocessed latent space of backbone neural network

Tomáš Chobola, Daniel Vašata, Pavel Kordík
AAAI Workshop on Meta-Learning and MetaDL Challenge, PMLR 140:29-37, 2021.

Abstract

The MetaDL Challenge 2020 focused on image classification tasks in few-shot settings. This paper describes second best submission in the competition. Our meta learning approach modifies the distribution of classes in a latent space produced by a backbone network for each class in order to better follow the Gaussian distribution. After this operation which we call Latent Space Transform algorithm, centers of classes are further aligned in an iterative fashion of the Expectation Maximisation algorithm to utilize information in unlabeled data that are often provided on top of few labelled instances. For this task, we utilize optimal transport mapping using the Sinkhorn algorithm. Our experiments show that this approach outperforms previous works as well as other variants of the algorithm, using K-Nearest Neighbour algorithm, Gaussian Mixture Models, etc.

Cite this Paper

BibTeX
@InProceedings{pmlr-v140-chobola21a, title = {Transfer learning based few-shot classification using optimal transport mapping from preprocessed latent space of backbone neural network}, author = {Chobola, Tom\'{a}\v{s} and Va\v{s}ata, Daniel and Kord\'{i}k, Pavel}, booktitle = {AAAI Workshop on Meta-Learning and MetaDL Challenge}, pages = {29--37}, 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/chobola21a/chobola21a.pdf}, url = {https://proceedings.mlr.press/v140/chobola21a.html}, abstract = {The MetaDL Challenge 2020 focused on image classification tasks in few-shot settings. This paper describes second best submission in the competition. Our meta learning approach modifies the distribution of classes in a latent space produced by a backbone network for each class in order to better follow the Gaussian distribution. After this operation which we call Latent Space Transform algorithm, centers of classes are further aligned in an iterative fashion of the Expectation Maximisation algorithm to utilize information in unlabeled data that are often provided on top of few labelled instances. For this task, we utilize optimal transport mapping using the Sinkhorn algorithm. Our experiments show that this approach outperforms previous works as well as other variants of the algorithm, using K-Nearest Neighbour algorithm, Gaussian Mixture Models, etc.} }
Endnote
%0 Conference Paper %T Transfer learning based few-shot classification using optimal transport mapping from preprocessed latent space of backbone neural network %A Tomáš Chobola %A Daniel Vašata %A Pavel Kordík %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-chobola21a %I PMLR %P 29--37 %U https://proceedings.mlr.press/v140/chobola21a.html %V 140 %X The MetaDL Challenge 2020 focused on image classification tasks in few-shot settings. This paper describes second best submission in the competition. Our meta learning approach modifies the distribution of classes in a latent space produced by a backbone network for each class in order to better follow the Gaussian distribution. After this operation which we call Latent Space Transform algorithm, centers of classes are further aligned in an iterative fashion of the Expectation Maximisation algorithm to utilize information in unlabeled data that are often provided on top of few labelled instances. For this task, we utilize optimal transport mapping using the Sinkhorn algorithm. Our experiments show that this approach outperforms previous works as well as other variants of the algorithm, using K-Nearest Neighbour algorithm, Gaussian Mixture Models, etc.
APA
Chobola, T., Vašata, D. & Kordík, P.. (2021). Transfer learning based few-shot classification using optimal transport mapping from preprocessed latent space of backbone neural network. AAAI Workshop on Meta-Learning and MetaDL Challenge, in Proceedings of Machine Learning Research 140:29-37 Available from https://proceedings.mlr.press/v140/chobola21a.html.

Related Material