Metadata Normalization
Batch Normalization (BN) and its variants have delivered tremendous success in combating the covariate shift induced by the training step of deep learning methods. While these techniques normalize feature distributions by standardizing with batch statistics, they do not correct the influence on feat...
Gespeichert in:
Veröffentlicht in: | arXiv.org 2021-05 |
---|---|
Hauptverfasser: | , , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | arXiv.org |
container_volume | |
creator | Lu, Mandy Zhao, Qingyu Zhang, Jiequan Pohl, Kilian M Li, Fei-Fei Niebles, Juan Carlos Adeli, Ehsan |
description | Batch Normalization (BN) and its variants have delivered tremendous success in combating the covariate shift induced by the training step of deep learning methods. While these techniques normalize feature distributions by standardizing with batch statistics, they do not correct the influence on features from extraneous variables or multiple distributions. Such extra variables, referred to as metadata here, may create bias or confounding effects (e.g., race when classifying gender from face images). We introduce the Metadata Normalization (MDN) layer, a new batch-level operation which can be used end-to-end within the training framework, to correct the influence of metadata on feature distributions. MDN adopts a regression analysis technique traditionally used for preprocessing to remove (regress out) the metadata effects on model features during training. We utilize a metric based on distance correlation to quantify the distribution bias from the metadata and demonstrate that our method successfully removes metadata effects on four diverse settings: one synthetic, one 2D image, one video, and one 3D medical image dataset. |
format | Article |
fullrecord | <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2515497117</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2515497117</sourcerecordid><originalsourceid>FETCH-proquest_journals_25154971173</originalsourceid><addsrcrecordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mQQ800tSUxJLElU8Msvyk3MyaxKLMnMz-NhYE1LzClO5YXS3AzKbq4hzh66BUX5haWpxSXxWfmlRXlAqXgjU0NTE0tzQ0NzY-JUAQAWqSmX</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2515497117</pqid></control><display><type>article</type><title>Metadata Normalization</title><source>Freely Accessible Journals</source><creator>Lu, Mandy ; Zhao, Qingyu ; Zhang, Jiequan ; Pohl, Kilian M ; Li, Fei-Fei ; Niebles, Juan Carlos ; Adeli, Ehsan</creator><creatorcontrib>Lu, Mandy ; Zhao, Qingyu ; Zhang, Jiequan ; Pohl, Kilian M ; Li, Fei-Fei ; Niebles, Juan Carlos ; Adeli, Ehsan</creatorcontrib><description>Batch Normalization (BN) and its variants have delivered tremendous success in combating the covariate shift induced by the training step of deep learning methods. While these techniques normalize feature distributions by standardizing with batch statistics, they do not correct the influence on features from extraneous variables or multiple distributions. Such extra variables, referred to as metadata here, may create bias or confounding effects (e.g., race when classifying gender from face images). We introduce the Metadata Normalization (MDN) layer, a new batch-level operation which can be used end-to-end within the training framework, to correct the influence of metadata on feature distributions. MDN adopts a regression analysis technique traditionally used for preprocessing to remove (regress out) the metadata effects on model features during training. We utilize a metric based on distance correlation to quantify the distribution bias from the metadata and demonstrate that our method successfully removes metadata effects on four diverse settings: one synthetic, one 2D image, one video, and one 3D medical image dataset.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Bias ; Image classification ; Machine learning ; Medical imaging ; Metadata ; Regression analysis ; Training</subject><ispartof>arXiv.org, 2021-05</ispartof><rights>2021. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>782,786</link.rule.ids></links><search><creatorcontrib>Lu, Mandy</creatorcontrib><creatorcontrib>Zhao, Qingyu</creatorcontrib><creatorcontrib>Zhang, Jiequan</creatorcontrib><creatorcontrib>Pohl, Kilian M</creatorcontrib><creatorcontrib>Li, Fei-Fei</creatorcontrib><creatorcontrib>Niebles, Juan Carlos</creatorcontrib><creatorcontrib>Adeli, Ehsan</creatorcontrib><title>Metadata Normalization</title><title>arXiv.org</title><description>Batch Normalization (BN) and its variants have delivered tremendous success in combating the covariate shift induced by the training step of deep learning methods. While these techniques normalize feature distributions by standardizing with batch statistics, they do not correct the influence on features from extraneous variables or multiple distributions. Such extra variables, referred to as metadata here, may create bias or confounding effects (e.g., race when classifying gender from face images). We introduce the Metadata Normalization (MDN) layer, a new batch-level operation which can be used end-to-end within the training framework, to correct the influence of metadata on feature distributions. MDN adopts a regression analysis technique traditionally used for preprocessing to remove (regress out) the metadata effects on model features during training. We utilize a metric based on distance correlation to quantify the distribution bias from the metadata and demonstrate that our method successfully removes metadata effects on four diverse settings: one synthetic, one 2D image, one video, and one 3D medical image dataset.</description><subject>Bias</subject><subject>Image classification</subject><subject>Machine learning</subject><subject>Medical imaging</subject><subject>Metadata</subject><subject>Regression analysis</subject><subject>Training</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><recordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mQQ800tSUxJLElU8Msvyk3MyaxKLMnMz-NhYE1LzClO5YXS3AzKbq4hzh66BUX5haWpxSXxWfmlRXlAqXgjU0NTE0tzQ0NzY-JUAQAWqSmX</recordid><startdate>20210505</startdate><enddate>20210505</enddate><creator>Lu, Mandy</creator><creator>Zhao, Qingyu</creator><creator>Zhang, Jiequan</creator><creator>Pohl, Kilian M</creator><creator>Li, Fei-Fei</creator><creator>Niebles, Juan Carlos</creator><creator>Adeli, Ehsan</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20210505</creationdate><title>Metadata Normalization</title><author>Lu, Mandy ; Zhao, Qingyu ; Zhang, Jiequan ; Pohl, Kilian M ; Li, Fei-Fei ; Niebles, Juan Carlos ; Adeli, Ehsan</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_25154971173</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Bias</topic><topic>Image classification</topic><topic>Machine learning</topic><topic>Medical imaging</topic><topic>Metadata</topic><topic>Regression analysis</topic><topic>Training</topic><toplevel>online_resources</toplevel><creatorcontrib>Lu, Mandy</creatorcontrib><creatorcontrib>Zhao, Qingyu</creatorcontrib><creatorcontrib>Zhang, Jiequan</creatorcontrib><creatorcontrib>Pohl, Kilian M</creatorcontrib><creatorcontrib>Li, Fei-Fei</creatorcontrib><creatorcontrib>Niebles, Juan Carlos</creatorcontrib><creatorcontrib>Adeli, Ehsan</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science & Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Access via ProQuest (Open Access)</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Lu, Mandy</au><au>Zhao, Qingyu</au><au>Zhang, Jiequan</au><au>Pohl, Kilian M</au><au>Li, Fei-Fei</au><au>Niebles, Juan Carlos</au><au>Adeli, Ehsan</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Metadata Normalization</atitle><jtitle>arXiv.org</jtitle><date>2021-05-05</date><risdate>2021</risdate><eissn>2331-8422</eissn><abstract>Batch Normalization (BN) and its variants have delivered tremendous success in combating the covariate shift induced by the training step of deep learning methods. While these techniques normalize feature distributions by standardizing with batch statistics, they do not correct the influence on features from extraneous variables or multiple distributions. Such extra variables, referred to as metadata here, may create bias or confounding effects (e.g., race when classifying gender from face images). We introduce the Metadata Normalization (MDN) layer, a new batch-level operation which can be used end-to-end within the training framework, to correct the influence of metadata on feature distributions. MDN adopts a regression analysis technique traditionally used for preprocessing to remove (regress out) the metadata effects on model features during training. We utilize a metric based on distance correlation to quantify the distribution bias from the metadata and demonstrate that our method successfully removes metadata effects on four diverse settings: one synthetic, one 2D image, one video, and one 3D medical image dataset.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | EISSN: 2331-8422 |
ispartof | arXiv.org, 2021-05 |
issn | 2331-8422 |
language | eng |
recordid | cdi_proquest_journals_2515497117 |
source | Freely Accessible Journals |
subjects | Bias Image classification Machine learning Medical imaging Metadata Regression analysis Training |
title | Metadata Normalization |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-03T01%3A49%3A18IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Metadata%20Normalization&rft.jtitle=arXiv.org&rft.au=Lu,%20Mandy&rft.date=2021-05-05&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2515497117%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2515497117&rft_id=info:pmid/&rfr_iscdi=true |