Unsupervised Metric Learning with Synthetic Examples
Distance Metric Learning (DML) involves learning an embedding that brings similar examples closer while moving away dissimilar ones. Existing DML approaches make use of class labels to generate constraints for metric learning. In this paper, we address the less-studied problem of learning a metric i...
Published in: | Proceedings of the AAAI Conference on Artificial Intelligence |
---|---|
Main Authors: | , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Association for the Advancement of Artificial Intelligence
2020
|
Subjects: | |
Online Access: | https://ojs.aaai.org/index.php/AAAI/article/view/5795 https://doi.org/10.1609/aaai.v34i04.5795 |
id |
ftjaaai:oai:ojs.aaai.org:article/5795 |
---|---|
record_format |
openpolar |
spelling |
ftjaaai:oai:ojs.aaai.org:article/5795 2023-05-15T16:01:25+02:00 Unsupervised Metric Learning with Synthetic Examples Dutta, Ujjal Kr Harandi, Mehrtash Sekhar, C. Chandra 2020-04-03 application/pdf https://ojs.aaai.org/index.php/AAAI/article/view/5795 https://doi.org/10.1609/aaai.v34i04.5795 eng eng Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AAAI/article/view/5795/5651 https://ojs.aaai.org/index.php/AAAI/article/view/5795 doi:10.1609/aaai.v34i04.5795 Copyright (c) 2020 Association for the Advancement of Artificial Intelligence Proceedings of the AAAI Conference on Artificial Intelligence; Vol. 34 No. 04: AAAI-20 Technical Tracks 4; 3834-3841 2374-3468 2159-5399 info:eu-repo/semantics/article info:eu-repo/semantics/publishedVersion 2020 ftjaaai https://doi.org/10.1609/aaai.v34i04.5795 2022-07-02T23:17:33Z Distance Metric Learning (DML) involves learning an embedding that brings similar examples closer while moving away dissimilar ones. Existing DML approaches make use of class labels to generate constraints for metric learning. In this paper, we address the less-studied problem of learning a metric in an unsupervised manner. We do not make use of class labels, but use unlabeled data to generate adversarial, synthetic constraints for learning a metric inducing embedding. Being a measure of uncertainty, we minimize the entropy of a conditional probability to learn the metric. Our stochastic formulation scales well to large datasets, and performs competitive to existing metric learning methods. Article in Journal/Newspaper DML AAAI Publications (Association for the Advancement of Artificial Intelligence) Proceedings of the AAAI Conference on Artificial Intelligence 34 04 3834 3841 |
institution |
Open Polar |
collection |
AAAI Publications (Association for the Advancement of Artificial Intelligence) |
op_collection_id |
ftjaaai |
language |
English |
description |
Distance Metric Learning (DML) involves learning an embedding that brings similar examples closer while moving away dissimilar ones. Existing DML approaches make use of class labels to generate constraints for metric learning. In this paper, we address the less-studied problem of learning a metric in an unsupervised manner. We do not make use of class labels, but use unlabeled data to generate adversarial, synthetic constraints for learning a metric inducing embedding. Being a measure of uncertainty, we minimize the entropy of a conditional probability to learn the metric. Our stochastic formulation scales well to large datasets, and performs competitive to existing metric learning methods. |
format |
Article in Journal/Newspaper |
author |
Dutta, Ujjal Kr Harandi, Mehrtash Sekhar, C. Chandra |
spellingShingle |
Dutta, Ujjal Kr Harandi, Mehrtash Sekhar, C. Chandra Unsupervised Metric Learning with Synthetic Examples |
author_facet |
Dutta, Ujjal Kr Harandi, Mehrtash Sekhar, C. Chandra |
author_sort |
Dutta, Ujjal Kr |
title |
Unsupervised Metric Learning with Synthetic Examples |
title_short |
Unsupervised Metric Learning with Synthetic Examples |
title_full |
Unsupervised Metric Learning with Synthetic Examples |
title_fullStr |
Unsupervised Metric Learning with Synthetic Examples |
title_full_unstemmed |
Unsupervised Metric Learning with Synthetic Examples |
title_sort |
unsupervised metric learning with synthetic examples |
publisher |
Association for the Advancement of Artificial Intelligence |
publishDate |
2020 |
url |
https://ojs.aaai.org/index.php/AAAI/article/view/5795 https://doi.org/10.1609/aaai.v34i04.5795 |
genre |
DML |
genre_facet |
DML |
op_source |
Proceedings of the AAAI Conference on Artificial Intelligence; Vol. 34 No. 04: AAAI-20 Technical Tracks 4; 3834-3841 2374-3468 2159-5399 |
op_relation |
https://ojs.aaai.org/index.php/AAAI/article/view/5795/5651 https://ojs.aaai.org/index.php/AAAI/article/view/5795 doi:10.1609/aaai.v34i04.5795 |
op_rights |
Copyright (c) 2020 Association for the Advancement of Artificial Intelligence |
op_doi |
https://doi.org/10.1609/aaai.v34i04.5795 |
container_title |
Proceedings of the AAAI Conference on Artificial Intelligence |
container_volume |
34 |
container_issue |
04 |
container_start_page |
3834 |
op_container_end_page |
3841 |
_version_ |
1766397286489784320 |