Human not in the loop: objective sample difficulty measures for Curriculum Learning

Curriculum learning is a learning method that trains models in a meaningful order from easier to harder samples. A key here is to devise automatic and objective difficulty measures of samples. In the medical domain, previous work applied domain knowledge from human experts to qualitatively assess cl...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhou, Zhengbo, Luo, Jun, Arefan, Dooman, Kitamura, Gene, Wu, Shandong
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Zhou, Zhengbo
Luo, Jun
Arefan, Dooman
Kitamura, Gene
Wu, Shandong
description Curriculum learning is a learning method that trains models in a meaningful order from easier to harder samples. A key here is to devise automatic and objective difficulty measures of samples. In the medical domain, previous work applied domain knowledge from human experts to qualitatively assess classification difficulty of medical images to guide curriculum learning, which requires extra annotation efforts, relies on subjective human experience, and may introduce bias. In this work, we propose a new automated curriculum learning technique using the variance of gradients (VoG) to compute an objective difficulty measure of samples and evaluated its effects on elbow fracture classification from X-ray images. Specifically, we used VoG as a metric to rank each sample in terms of the classification difficulty, where high VoG scores indicate more difficult cases for classification, to guide the curriculum training process We compared the proposed technique to a baseline (without curriculum learning), a previous method that used human annotations on classification difficulty, and anti-curriculum learning. Our experiment results showed comparable and higher performance for the binary and multi-class bone fracture classification tasks.
doi_str_mv 10.48550/arxiv.2302.01243
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2302_01243</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2302_01243</sourcerecordid><originalsourceid>FETCH-LOGICAL-a673-f12cc01660e31f3a6a6f49a7e59062cc1fb569fc9ec207ba0e0f1f657d58e7883</originalsourceid><addsrcrecordid>eNotz81OwzAQBGBfOKDCA3BiXyDBjmMn4YYioEiReqD3aOPuglH-5MQVfXto4TTSjDTSJ8SdkmleGiMfMHz7Y5ppmaVSZbm-Fu_bOOAI47SCH2H9JOinaX6Eqfsit_ojwYLD3BMcPLN3sV9PMBAuMdACPAWoYwjnPg7QEIbRjx834oqxX-j2Pzdi__K8r7dJs3t9q5-aBG2hE1aZc1JZK0kr1mjRcl5hQaaS9ndS3BlbsavIZbLoUJJkxdYUB1NSUZZ6I-7_bi-qdg5-wHBqz7r2otM_eSNLIg</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Human not in the loop: objective sample difficulty measures for Curriculum Learning</title><source>arXiv.org</source><creator>Zhou, Zhengbo ; Luo, Jun ; Arefan, Dooman ; Kitamura, Gene ; Wu, Shandong</creator><creatorcontrib>Zhou, Zhengbo ; Luo, Jun ; Arefan, Dooman ; Kitamura, Gene ; Wu, Shandong</creatorcontrib><description>Curriculum learning is a learning method that trains models in a meaningful order from easier to harder samples. A key here is to devise automatic and objective difficulty measures of samples. In the medical domain, previous work applied domain knowledge from human experts to qualitatively assess classification difficulty of medical images to guide curriculum learning, which requires extra annotation efforts, relies on subjective human experience, and may introduce bias. In this work, we propose a new automated curriculum learning technique using the variance of gradients (VoG) to compute an objective difficulty measure of samples and evaluated its effects on elbow fracture classification from X-ray images. Specifically, we used VoG as a metric to rank each sample in terms of the classification difficulty, where high VoG scores indicate more difficult cases for classification, to guide the curriculum training process We compared the proposed technique to a baseline (without curriculum learning), a previous method that used human annotations on classification difficulty, and anti-curriculum learning. Our experiment results showed comparable and higher performance for the binary and multi-class bone fracture classification tasks.</description><identifier>DOI: 10.48550/arxiv.2302.01243</identifier><language>eng</language><subject>Computer Science - Computer Vision and Pattern Recognition ; Computer Science - Learning</subject><creationdate>2023-02</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2302.01243$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2302.01243$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Zhou, Zhengbo</creatorcontrib><creatorcontrib>Luo, Jun</creatorcontrib><creatorcontrib>Arefan, Dooman</creatorcontrib><creatorcontrib>Kitamura, Gene</creatorcontrib><creatorcontrib>Wu, Shandong</creatorcontrib><title>Human not in the loop: objective sample difficulty measures for Curriculum Learning</title><description>Curriculum learning is a learning method that trains models in a meaningful order from easier to harder samples. A key here is to devise automatic and objective difficulty measures of samples. In the medical domain, previous work applied domain knowledge from human experts to qualitatively assess classification difficulty of medical images to guide curriculum learning, which requires extra annotation efforts, relies on subjective human experience, and may introduce bias. In this work, we propose a new automated curriculum learning technique using the variance of gradients (VoG) to compute an objective difficulty measure of samples and evaluated its effects on elbow fracture classification from X-ray images. Specifically, we used VoG as a metric to rank each sample in terms of the classification difficulty, where high VoG scores indicate more difficult cases for classification, to guide the curriculum training process We compared the proposed technique to a baseline (without curriculum learning), a previous method that used human annotations on classification difficulty, and anti-curriculum learning. Our experiment results showed comparable and higher performance for the binary and multi-class bone fracture classification tasks.</description><subject>Computer Science - Computer Vision and Pattern Recognition</subject><subject>Computer Science - Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotz81OwzAQBGBfOKDCA3BiXyDBjmMn4YYioEiReqD3aOPuglH-5MQVfXto4TTSjDTSJ8SdkmleGiMfMHz7Y5ppmaVSZbm-Fu_bOOAI47SCH2H9JOinaX6Eqfsit_ojwYLD3BMcPLN3sV9PMBAuMdACPAWoYwjnPg7QEIbRjx834oqxX-j2Pzdi__K8r7dJs3t9q5-aBG2hE1aZc1JZK0kr1mjRcl5hQaaS9ndS3BlbsavIZbLoUJJkxdYUB1NSUZZ6I-7_bi-qdg5-wHBqz7r2otM_eSNLIg</recordid><startdate>20230202</startdate><enddate>20230202</enddate><creator>Zhou, Zhengbo</creator><creator>Luo, Jun</creator><creator>Arefan, Dooman</creator><creator>Kitamura, Gene</creator><creator>Wu, Shandong</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20230202</creationdate><title>Human not in the loop: objective sample difficulty measures for Curriculum Learning</title><author>Zhou, Zhengbo ; Luo, Jun ; Arefan, Dooman ; Kitamura, Gene ; Wu, Shandong</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a673-f12cc01660e31f3a6a6f49a7e59062cc1fb569fc9ec207ba0e0f1f657d58e7883</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Computer Science - Computer Vision and Pattern Recognition</topic><topic>Computer Science - Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Zhou, Zhengbo</creatorcontrib><creatorcontrib>Luo, Jun</creatorcontrib><creatorcontrib>Arefan, Dooman</creatorcontrib><creatorcontrib>Kitamura, Gene</creatorcontrib><creatorcontrib>Wu, Shandong</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Zhou, Zhengbo</au><au>Luo, Jun</au><au>Arefan, Dooman</au><au>Kitamura, Gene</au><au>Wu, Shandong</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Human not in the loop: objective sample difficulty measures for Curriculum Learning</atitle><date>2023-02-02</date><risdate>2023</risdate><abstract>Curriculum learning is a learning method that trains models in a meaningful order from easier to harder samples. A key here is to devise automatic and objective difficulty measures of samples. In the medical domain, previous work applied domain knowledge from human experts to qualitatively assess classification difficulty of medical images to guide curriculum learning, which requires extra annotation efforts, relies on subjective human experience, and may introduce bias. In this work, we propose a new automated curriculum learning technique using the variance of gradients (VoG) to compute an objective difficulty measure of samples and evaluated its effects on elbow fracture classification from X-ray images. Specifically, we used VoG as a metric to rank each sample in terms of the classification difficulty, where high VoG scores indicate more difficult cases for classification, to guide the curriculum training process We compared the proposed technique to a baseline (without curriculum learning), a previous method that used human annotations on classification difficulty, and anti-curriculum learning. Our experiment results showed comparable and higher performance for the binary and multi-class bone fracture classification tasks.</abstract><doi>10.48550/arxiv.2302.01243</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2302.01243
ispartof
issn
language eng
recordid cdi_arxiv_primary_2302_01243
source arXiv.org
subjects Computer Science - Computer Vision and Pattern Recognition
Computer Science - Learning
title Human not in the loop: objective sample difficulty measures for Curriculum Learning
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-12T08%3A43%3A49IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Human%20not%20in%20the%20loop:%20objective%20sample%20difficulty%20measures%20for%20Curriculum%20Learning&rft.au=Zhou,%20Zhengbo&rft.date=2023-02-02&rft_id=info:doi/10.48550/arxiv.2302.01243&rft_dat=%3Carxiv_GOX%3E2302_01243%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true