MILES: Multiple-Instance Learning via Embedded Instance Selection
Multiple-instance problems arise from the situations where training class labels are attached to sets of samples (named bags), instead of individual samples within each bag (called instances). Most previous multiple-instance learning (MIL) algorithms are developed based on the assumption that a bag...
Gespeichert in:
Veröffentlicht in: | IEEE transactions on pattern analysis and machine intelligence 2006-12, Vol.28 (12), p.1931-1947 |
---|---|
Hauptverfasser: | , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | 1947 |
---|---|
container_issue | 12 |
container_start_page | 1931 |
container_title | IEEE transactions on pattern analysis and machine intelligence |
container_volume | 28 |
creator | Yixin Chen Jinbo Bi Wang, J.Z. |
description | Multiple-instance problems arise from the situations where training class labels are attached to sets of samples (named bags), instead of individual samples within each bag (called instances). Most previous multiple-instance learning (MIL) algorithms are developed based on the assumption that a bag is positive if and only if at least one of its instances is positive. Although the assumption works well in a drug activity prediction problem, it is rather restrictive for other applications, especially those in the computer vision area. We propose a learning method, MILES (multiple-instance learning via embedded instance selection), which converts the multiple-instance learning problem to a standard supervised learning problem that does not impose the assumption relating instance labels to bag labels. MILES maps each bag into a feature space defined by the instances in the training bags via an instance similarity measure. This feature mapping often provides a large number of redundant or irrelevant features. Hence, 1-norm SVM is applied to select important features as well as construct classifiers simultaneously. We have performed extensive experiments. In comparison with other methods, MILES demonstrates competitive classification accuracy, high computation efficiency, and robustness to labeling uncertainty |
doi_str_mv | 10.1109/TPAMI.2006.248 |
format | Article |
fullrecord | <record><control><sourceid>proquest_RIE</sourceid><recordid>TN_cdi_pubmed_primary_17108368</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>1717454</ieee_id><sourcerecordid>896193362</sourcerecordid><originalsourceid>FETCH-LOGICAL-c402t-fd8b2c733ca5fb7adf58cd111a213cb56ac0cfea9f2eb82ab6cc543076af44c23</originalsourceid><addsrcrecordid>eNp90c9LwzAUB_Agips_rl4EKYJ66sxL2jT1NmTqYKKweQ5p-iIdXTubVvC_N3VDxYOnHN7nvfDel5AToCMAml4vnseP0xGjVIxYJHfIEFKehjzm6S4ZUhAslJLJATlwbkkpRDHl-2QACVDJhRwS3z2bzG-Cx65si3WJ4bRyra4MBjPUTVVUr8F7oYPJKsM8xzz4Ls-xRNMWdXVE9qwuHR5v30PycjdZ3D6Es6f76e14FpqIsja0ucyYSTg3OrZZonMbS5MDgGbATRYLbaixqFPLMJNMZ8KYOOI0EdpGkWH8kFxt5q6b-q1D16pV4QyWpa6w7pySqfC7c9HLy3-lkBBz4D08_wOXdddUfgslRZwASKAejTbINLVzDVq1boqVbj4UUNVnoL4yUH0GymfgG862U7tshfkP3x7dg4st0M7o0jb-oIX7cZL5ryPh3enGFYj4e0wS-ct8AruyliY</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>865711810</pqid></control><display><type>article</type><title>MILES: Multiple-Instance Learning via Embedded Instance Selection</title><source>IEEE Electronic Library Online</source><creator>Yixin Chen ; Jinbo Bi ; Wang, J.Z.</creator><creatorcontrib>Yixin Chen ; Jinbo Bi ; Wang, J.Z.</creatorcontrib><description>Multiple-instance problems arise from the situations where training class labels are attached to sets of samples (named bags), instead of individual samples within each bag (called instances). Most previous multiple-instance learning (MIL) algorithms are developed based on the assumption that a bag is positive if and only if at least one of its instances is positive. Although the assumption works well in a drug activity prediction problem, it is rather restrictive for other applications, especially those in the computer vision area. We propose a learning method, MILES (multiple-instance learning via embedded instance selection), which converts the multiple-instance learning problem to a standard supervised learning problem that does not impose the assumption relating instance labels to bag labels. MILES maps each bag into a feature space defined by the instances in the training bags via an instance similarity measure. This feature mapping often provides a large number of redundant or irrelevant features. Hence, 1-norm SVM is applied to select important features as well as construct classifiers simultaneously. We have performed extensive experiments. In comparison with other methods, MILES demonstrates competitive classification accuracy, high computation efficiency, and robustness to labeling uncertainty</description><identifier>ISSN: 0162-8828</identifier><identifier>EISSN: 1939-3539</identifier><identifier>EISSN: 2160-9292</identifier><identifier>DOI: 10.1109/TPAMI.2006.248</identifier><identifier>PMID: 17108368</identifier><identifier>CODEN: ITPIDJ</identifier><language>eng</language><publisher>Los Alamitos, CA: IEEE</publisher><subject>1-norm support vector machine ; Algorithms ; Application software ; Applied sciences ; Artificial Intelligence ; Bags ; Computer science; control theory; systems ; Computer vision ; drug activity prediction ; Drugs ; Exact sciences and technology ; feature subset selection ; image categorization ; Image Enhancement - methods ; Image Interpretation, Computer-Assisted - methods ; Imaging, Three-Dimensional - methods ; Information Storage and Retrieval - methods ; Intelligence ; Labeling ; Labels ; Learning ; Learning systems ; Multiple-instance learning ; object recognition ; Pattern analysis ; Pattern recognition. Digital image processing. Computational geometry ; Reproducibility of Results ; Robustness ; Sensitivity and Specificity ; Similarity ; Studies ; Supervised learning ; Support vector machine classification ; Support vector machines ; Training ; Uncertainty</subject><ispartof>IEEE transactions on pattern analysis and machine intelligence, 2006-12, Vol.28 (12), p.1931-1947</ispartof><rights>2007 INIST-CNRS</rights><rights>Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2006</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c402t-fd8b2c733ca5fb7adf58cd111a213cb56ac0cfea9f2eb82ab6cc543076af44c23</citedby><cites>FETCH-LOGICAL-c402t-fd8b2c733ca5fb7adf58cd111a213cb56ac0cfea9f2eb82ab6cc543076af44c23</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/1717454$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,780,784,796,27924,27925,54758</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/1717454$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc><backlink>$$Uhttp://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=18271146$$DView record in Pascal Francis$$Hfree_for_read</backlink><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/17108368$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Yixin Chen</creatorcontrib><creatorcontrib>Jinbo Bi</creatorcontrib><creatorcontrib>Wang, J.Z.</creatorcontrib><title>MILES: Multiple-Instance Learning via Embedded Instance Selection</title><title>IEEE transactions on pattern analysis and machine intelligence</title><addtitle>TPAMI</addtitle><addtitle>IEEE Trans Pattern Anal Mach Intell</addtitle><description>Multiple-instance problems arise from the situations where training class labels are attached to sets of samples (named bags), instead of individual samples within each bag (called instances). Most previous multiple-instance learning (MIL) algorithms are developed based on the assumption that a bag is positive if and only if at least one of its instances is positive. Although the assumption works well in a drug activity prediction problem, it is rather restrictive for other applications, especially those in the computer vision area. We propose a learning method, MILES (multiple-instance learning via embedded instance selection), which converts the multiple-instance learning problem to a standard supervised learning problem that does not impose the assumption relating instance labels to bag labels. MILES maps each bag into a feature space defined by the instances in the training bags via an instance similarity measure. This feature mapping often provides a large number of redundant or irrelevant features. Hence, 1-norm SVM is applied to select important features as well as construct classifiers simultaneously. We have performed extensive experiments. In comparison with other methods, MILES demonstrates competitive classification accuracy, high computation efficiency, and robustness to labeling uncertainty</description><subject>1-norm support vector machine</subject><subject>Algorithms</subject><subject>Application software</subject><subject>Applied sciences</subject><subject>Artificial Intelligence</subject><subject>Bags</subject><subject>Computer science; control theory; systems</subject><subject>Computer vision</subject><subject>drug activity prediction</subject><subject>Drugs</subject><subject>Exact sciences and technology</subject><subject>feature subset selection</subject><subject>image categorization</subject><subject>Image Enhancement - methods</subject><subject>Image Interpretation, Computer-Assisted - methods</subject><subject>Imaging, Three-Dimensional - methods</subject><subject>Information Storage and Retrieval - methods</subject><subject>Intelligence</subject><subject>Labeling</subject><subject>Labels</subject><subject>Learning</subject><subject>Learning systems</subject><subject>Multiple-instance learning</subject><subject>object recognition</subject><subject>Pattern analysis</subject><subject>Pattern recognition. Digital image processing. Computational geometry</subject><subject>Reproducibility of Results</subject><subject>Robustness</subject><subject>Sensitivity and Specificity</subject><subject>Similarity</subject><subject>Studies</subject><subject>Supervised learning</subject><subject>Support vector machine classification</subject><subject>Support vector machines</subject><subject>Training</subject><subject>Uncertainty</subject><issn>0162-8828</issn><issn>1939-3539</issn><issn>2160-9292</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2006</creationdate><recordtype>article</recordtype><sourceid>RIE</sourceid><sourceid>EIF</sourceid><recordid>eNp90c9LwzAUB_Agips_rl4EKYJ66sxL2jT1NmTqYKKweQ5p-iIdXTubVvC_N3VDxYOnHN7nvfDel5AToCMAml4vnseP0xGjVIxYJHfIEFKehjzm6S4ZUhAslJLJATlwbkkpRDHl-2QACVDJhRwS3z2bzG-Cx65si3WJ4bRyra4MBjPUTVVUr8F7oYPJKsM8xzz4Ls-xRNMWdXVE9qwuHR5v30PycjdZ3D6Es6f76e14FpqIsja0ucyYSTg3OrZZonMbS5MDgGbATRYLbaixqFPLMJNMZ8KYOOI0EdpGkWH8kFxt5q6b-q1D16pV4QyWpa6w7pySqfC7c9HLy3-lkBBz4D08_wOXdddUfgslRZwASKAejTbINLVzDVq1boqVbj4UUNVnoL4yUH0GymfgG862U7tshfkP3x7dg4st0M7o0jb-oIX7cZL5ryPh3enGFYj4e0wS-ct8AruyliY</recordid><startdate>20061201</startdate><enddate>20061201</enddate><creator>Yixin Chen</creator><creator>Jinbo Bi</creator><creator>Wang, J.Z.</creator><general>IEEE</general><general>IEEE Computer Society</general><general>The Institute of Electrical and Electronics Engineers, Inc. (IEEE)</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>IQODW</scope><scope>CGR</scope><scope>CUY</scope><scope>CVF</scope><scope>ECM</scope><scope>EIF</scope><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>7SP</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><scope>7X8</scope><scope>F28</scope><scope>FR3</scope></search><sort><creationdate>20061201</creationdate><title>MILES: Multiple-Instance Learning via Embedded Instance Selection</title><author>Yixin Chen ; Jinbo Bi ; Wang, J.Z.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c402t-fd8b2c733ca5fb7adf58cd111a213cb56ac0cfea9f2eb82ab6cc543076af44c23</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2006</creationdate><topic>1-norm support vector machine</topic><topic>Algorithms</topic><topic>Application software</topic><topic>Applied sciences</topic><topic>Artificial Intelligence</topic><topic>Bags</topic><topic>Computer science; control theory; systems</topic><topic>Computer vision</topic><topic>drug activity prediction</topic><topic>Drugs</topic><topic>Exact sciences and technology</topic><topic>feature subset selection</topic><topic>image categorization</topic><topic>Image Enhancement - methods</topic><topic>Image Interpretation, Computer-Assisted - methods</topic><topic>Imaging, Three-Dimensional - methods</topic><topic>Information Storage and Retrieval - methods</topic><topic>Intelligence</topic><topic>Labeling</topic><topic>Labels</topic><topic>Learning</topic><topic>Learning systems</topic><topic>Multiple-instance learning</topic><topic>object recognition</topic><topic>Pattern analysis</topic><topic>Pattern recognition. Digital image processing. Computational geometry</topic><topic>Reproducibility of Results</topic><topic>Robustness</topic><topic>Sensitivity and Specificity</topic><topic>Similarity</topic><topic>Studies</topic><topic>Supervised learning</topic><topic>Support vector machine classification</topic><topic>Support vector machines</topic><topic>Training</topic><topic>Uncertainty</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Yixin Chen</creatorcontrib><creatorcontrib>Jinbo Bi</creatorcontrib><creatorcontrib>Wang, J.Z.</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998–Present</collection><collection>IEEE Electronic Library Online</collection><collection>Pascal-Francis</collection><collection>Medline</collection><collection>MEDLINE</collection><collection>MEDLINE (Ovid)</collection><collection>MEDLINE</collection><collection>MEDLINE</collection><collection>PubMed</collection><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Electronics & Communications Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><collection>MEDLINE - Academic</collection><collection>ANTE: Abstracts in New Technology & Engineering</collection><collection>Engineering Research Database</collection><jtitle>IEEE transactions on pattern analysis and machine intelligence</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Yixin Chen</au><au>Jinbo Bi</au><au>Wang, J.Z.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>MILES: Multiple-Instance Learning via Embedded Instance Selection</atitle><jtitle>IEEE transactions on pattern analysis and machine intelligence</jtitle><stitle>TPAMI</stitle><addtitle>IEEE Trans Pattern Anal Mach Intell</addtitle><date>2006-12-01</date><risdate>2006</risdate><volume>28</volume><issue>12</issue><spage>1931</spage><epage>1947</epage><pages>1931-1947</pages><issn>0162-8828</issn><eissn>1939-3539</eissn><eissn>2160-9292</eissn><coden>ITPIDJ</coden><abstract>Multiple-instance problems arise from the situations where training class labels are attached to sets of samples (named bags), instead of individual samples within each bag (called instances). Most previous multiple-instance learning (MIL) algorithms are developed based on the assumption that a bag is positive if and only if at least one of its instances is positive. Although the assumption works well in a drug activity prediction problem, it is rather restrictive for other applications, especially those in the computer vision area. We propose a learning method, MILES (multiple-instance learning via embedded instance selection), which converts the multiple-instance learning problem to a standard supervised learning problem that does not impose the assumption relating instance labels to bag labels. MILES maps each bag into a feature space defined by the instances in the training bags via an instance similarity measure. This feature mapping often provides a large number of redundant or irrelevant features. Hence, 1-norm SVM is applied to select important features as well as construct classifiers simultaneously. We have performed extensive experiments. In comparison with other methods, MILES demonstrates competitive classification accuracy, high computation efficiency, and robustness to labeling uncertainty</abstract><cop>Los Alamitos, CA</cop><pub>IEEE</pub><pmid>17108368</pmid><doi>10.1109/TPAMI.2006.248</doi><tpages>17</tpages></addata></record> |
fulltext | fulltext_linktorsrc |
identifier | ISSN: 0162-8828 |
ispartof | IEEE transactions on pattern analysis and machine intelligence, 2006-12, Vol.28 (12), p.1931-1947 |
issn | 0162-8828 1939-3539 2160-9292 |
language | eng |
recordid | cdi_pubmed_primary_17108368 |
source | IEEE Electronic Library Online |
subjects | 1-norm support vector machine Algorithms Application software Applied sciences Artificial Intelligence Bags Computer science control theory systems Computer vision drug activity prediction Drugs Exact sciences and technology feature subset selection image categorization Image Enhancement - methods Image Interpretation, Computer-Assisted - methods Imaging, Three-Dimensional - methods Information Storage and Retrieval - methods Intelligence Labeling Labels Learning Learning systems Multiple-instance learning object recognition Pattern analysis Pattern recognition. Digital image processing. Computational geometry Reproducibility of Results Robustness Sensitivity and Specificity Similarity Studies Supervised learning Support vector machine classification Support vector machines Training Uncertainty |
title | MILES: Multiple-Instance Learning via Embedded Instance Selection |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-25T06%3A20%3A11IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=MILES:%20Multiple-Instance%20Learning%20via%20Embedded%20Instance%20Selection&rft.jtitle=IEEE%20transactions%20on%20pattern%20analysis%20and%20machine%20intelligence&rft.au=Yixin%20Chen&rft.date=2006-12-01&rft.volume=28&rft.issue=12&rft.spage=1931&rft.epage=1947&rft.pages=1931-1947&rft.issn=0162-8828&rft.eissn=1939-3539&rft.coden=ITPIDJ&rft_id=info:doi/10.1109/TPAMI.2006.248&rft_dat=%3Cproquest_RIE%3E896193362%3C/proquest_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=865711810&rft_id=info:pmid/17108368&rft_ieee_id=1717454&rfr_iscdi=true |