Self-Supervised Knowledge Transfer via Loosely Supervised Auxiliary Tasks
Knowledge transfer using convolutional neural networks (CNNs) can help efficiently train a CNN with fewer parameters or maximize the generalization performance under limited supervision. To enable a more efficient transfer of pretrained knowledge under relaxed conditions, we propose a simple yet pow...
Main Authors: | , , , |
---|---|
Format: | Article in Journal/Newspaper |
Language: | unknown |
Published: |
arXiv
2021
|
Subjects: | |
Online Access: | https://dx.doi.org/10.48550/arxiv.2110.12696 https://arxiv.org/abs/2110.12696 |
id |
ftdatacite:10.48550/arxiv.2110.12696 |
---|---|
record_format |
openpolar |
spelling |
ftdatacite:10.48550/arxiv.2110.12696 2023-05-15T16:02:00+02:00 Self-Supervised Knowledge Transfer via Loosely Supervised Auxiliary Tasks Hong, Seungbum Yoon, Jihun Kim, Junmo Choi, Min-Kook 2021 https://dx.doi.org/10.48550/arxiv.2110.12696 https://arxiv.org/abs/2110.12696 unknown arXiv Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode cc-by-4.0 CC-BY Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences Article CreativeWork article Preprint 2021 ftdatacite https://doi.org/10.48550/arxiv.2110.12696 2022-03-10T14:00:00Z Knowledge transfer using convolutional neural networks (CNNs) can help efficiently train a CNN with fewer parameters or maximize the generalization performance under limited supervision. To enable a more efficient transfer of pretrained knowledge under relaxed conditions, we propose a simple yet powerful knowledge transfer methodology without any restrictions regarding the network structure or dataset used, namely self-supervised knowledge transfer (SSKT), via loosely supervised auxiliary tasks. For this, we devise a training methodology that transfers previously learned knowledge to the current training process as an auxiliary task for the target task through self-supervision using a soft label. The SSKT is independent of the network structure and dataset, and is trained differently from existing knowledge transfer methods; hence, it has an advantage in that the prior knowledge acquired from various tasks can be naturally transferred during the training process to the target task. Furthermore, it can improve the generalization performance on most datasets through the proposed knowledge transfer between different problem domains from multiple source networks. SSKT outperforms the other transfer learning methods (KD, DML, and MAXL) through experiments under various knowledge transfer settings. The source code will be made available to the public. : Accepted at WACV 2022 Article in Journal/Newspaper DML DataCite Metadata Store (German National Library of Science and Technology) |
institution |
Open Polar |
collection |
DataCite Metadata Store (German National Library of Science and Technology) |
op_collection_id |
ftdatacite |
language |
unknown |
topic |
Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences |
spellingShingle |
Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences Hong, Seungbum Yoon, Jihun Kim, Junmo Choi, Min-Kook Self-Supervised Knowledge Transfer via Loosely Supervised Auxiliary Tasks |
topic_facet |
Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences |
description |
Knowledge transfer using convolutional neural networks (CNNs) can help efficiently train a CNN with fewer parameters or maximize the generalization performance under limited supervision. To enable a more efficient transfer of pretrained knowledge under relaxed conditions, we propose a simple yet powerful knowledge transfer methodology without any restrictions regarding the network structure or dataset used, namely self-supervised knowledge transfer (SSKT), via loosely supervised auxiliary tasks. For this, we devise a training methodology that transfers previously learned knowledge to the current training process as an auxiliary task for the target task through self-supervision using a soft label. The SSKT is independent of the network structure and dataset, and is trained differently from existing knowledge transfer methods; hence, it has an advantage in that the prior knowledge acquired from various tasks can be naturally transferred during the training process to the target task. Furthermore, it can improve the generalization performance on most datasets through the proposed knowledge transfer between different problem domains from multiple source networks. SSKT outperforms the other transfer learning methods (KD, DML, and MAXL) through experiments under various knowledge transfer settings. The source code will be made available to the public. : Accepted at WACV 2022 |
format |
Article in Journal/Newspaper |
author |
Hong, Seungbum Yoon, Jihun Kim, Junmo Choi, Min-Kook |
author_facet |
Hong, Seungbum Yoon, Jihun Kim, Junmo Choi, Min-Kook |
author_sort |
Hong, Seungbum |
title |
Self-Supervised Knowledge Transfer via Loosely Supervised Auxiliary Tasks |
title_short |
Self-Supervised Knowledge Transfer via Loosely Supervised Auxiliary Tasks |
title_full |
Self-Supervised Knowledge Transfer via Loosely Supervised Auxiliary Tasks |
title_fullStr |
Self-Supervised Knowledge Transfer via Loosely Supervised Auxiliary Tasks |
title_full_unstemmed |
Self-Supervised Knowledge Transfer via Loosely Supervised Auxiliary Tasks |
title_sort |
self-supervised knowledge transfer via loosely supervised auxiliary tasks |
publisher |
arXiv |
publishDate |
2021 |
url |
https://dx.doi.org/10.48550/arxiv.2110.12696 https://arxiv.org/abs/2110.12696 |
genre |
DML |
genre_facet |
DML |
op_rights |
Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode cc-by-4.0 |
op_rightsnorm |
CC-BY |
op_doi |
https://doi.org/10.48550/arxiv.2110.12696 |
_version_ |
1766397647490383872 |