Loading…
Hyperbolic Secant representation of the logistic function: Application to probabilistic Multiple Instance Learning for CT intracranial hemorrhage detection
Multiple Instance Learning (MIL) is a weakly supervised paradigm that has been successfully applied to many different scientific areas and is particularly well suited to medical imaging. Probabilistic MIL methods, and more specifically Gaussian Processes (GPs), have achieved excellent results due to...
Saved in:
Published in: | Artificial intelligence 2024-06, Vol.331, p.104115, Article 104115 |
---|---|
Main Authors: | , , , , |
Format: | Article |
Language: | English |
Subjects: | |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
cited_by | cdi_FETCH-LOGICAL-c352t-707041ac70c3f7b5595b2ef0940fa4eb5f1b0e7e17324979dd93fd83687d185e3 |
---|---|
cites | cdi_FETCH-LOGICAL-c352t-707041ac70c3f7b5595b2ef0940fa4eb5f1b0e7e17324979dd93fd83687d185e3 |
container_end_page | |
container_issue | |
container_start_page | 104115 |
container_title | Artificial intelligence |
container_volume | 331 |
creator | Castro-Macías, Francisco M. Morales-Álvarez, Pablo Wu, Yunan Molina, Rafael Katsaggelos, Aggelos K. |
description | Multiple Instance Learning (MIL) is a weakly supervised paradigm that has been successfully applied to many different scientific areas and is particularly well suited to medical imaging. Probabilistic MIL methods, and more specifically Gaussian Processes (GPs), have achieved excellent results due to their high expressiveness and uncertainty quantification capabilities. One of the most successful GP-based MIL methods, VGPMIL, resorts to a variational bound to handle the intractability of the logistic function. Here, we formulate VGPMIL using Pólya-Gamma random variables. This approach yields the same variational posterior approximations as the original VGPMIL, which is a consequence of the two representations that the Hyperbolic Secant distribution admits. This leads us to propose a general GP-based MIL method that takes different forms by simply leveraging distributions other than the Hyperbolic Secant one. Using the Gamma distribution we arrive at a new approach that obtains competitive or superior predictive performance and efficiency. This is validated in a comprehensive experimental study including one synthetic MIL dataset, two well-known MIL benchmarks, and a real-world medical problem. We expect that this work provides useful ideas beyond MIL that can foster further research in the field. |
doi_str_mv | 10.1016/j.artint.2024.104115 |
format | article |
fullrecord | <record><control><sourceid>elsevier_cross</sourceid><recordid>TN_cdi_crossref_primary_10_1016_j_artint_2024_104115</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><els_id>S0004370224000511</els_id><sourcerecordid>S0004370224000511</sourcerecordid><originalsourceid>FETCH-LOGICAL-c352t-707041ac70c3f7b5595b2ef0940fa4eb5f1b0e7e17324979dd93fd83687d185e3</originalsourceid><addsrcrecordid>eNp9UMtOwzAQtBBIlMIfcPAPpNh51A0HpKoCWqmIA-VsOc66dZXa0dpF6rfws7iEM6fV7s7szgwh95xNOOPTh_1EYbQuTnKWl2lUcl5dkBGfiTwTdc4vyYgxVmaFYPk1uQlhn9qirvmIfC9PPWDjO6vpB2jlIkXoEQK4qKL1jnpD4w5o57c2xIQyR6fPi0c67_tEG1DR0x59oxrbDbC3Yxdt3wFduRCV00DXoNBZt6XGI11saBKMSqNyVnV0BwePuFNboC1E-P1wS66M6gLc_dUx-Xx53iyW2fr9dbWYrzNdVHnMBBPJsdKC6cKIpqrqqsnBsLpkRpXQVIY3DARwUeRlLeq2rQvTzorpTLR8VkExJuVwV6MPAcHIHu1B4UlyJs8By70cApbngOUQcKI9DTRI2r4soAzaQnLaWkwGZOvt_wd-AMulioY</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Hyperbolic Secant representation of the logistic function: Application to probabilistic Multiple Instance Learning for CT intracranial hemorrhage detection</title><source>Elsevier:Jisc Collections:Elsevier Read and Publish Agreement 2022-2024:Freedom Collection (Reading list)</source><creator>Castro-Macías, Francisco M. ; Morales-Álvarez, Pablo ; Wu, Yunan ; Molina, Rafael ; Katsaggelos, Aggelos K.</creator><creatorcontrib>Castro-Macías, Francisco M. ; Morales-Álvarez, Pablo ; Wu, Yunan ; Molina, Rafael ; Katsaggelos, Aggelos K.</creatorcontrib><description>Multiple Instance Learning (MIL) is a weakly supervised paradigm that has been successfully applied to many different scientific areas and is particularly well suited to medical imaging. Probabilistic MIL methods, and more specifically Gaussian Processes (GPs), have achieved excellent results due to their high expressiveness and uncertainty quantification capabilities. One of the most successful GP-based MIL methods, VGPMIL, resorts to a variational bound to handle the intractability of the logistic function. Here, we formulate VGPMIL using Pólya-Gamma random variables. This approach yields the same variational posterior approximations as the original VGPMIL, which is a consequence of the two representations that the Hyperbolic Secant distribution admits. This leads us to propose a general GP-based MIL method that takes different forms by simply leveraging distributions other than the Hyperbolic Secant one. Using the Gamma distribution we arrive at a new approach that obtains competitive or superior predictive performance and efficiency. This is validated in a comprehensive experimental study including one synthetic MIL dataset, two well-known MIL benchmarks, and a real-world medical problem. We expect that this work provides useful ideas beyond MIL that can foster further research in the field.</description><identifier>ISSN: 0004-3702</identifier><identifier>EISSN: 1872-7921</identifier><identifier>DOI: 10.1016/j.artint.2024.104115</identifier><language>eng</language><publisher>Elsevier B.V</publisher><subject>Gaussian processes ; Hyperbolic Secant distribution ; Intracranial hemorrhage detection ; Jaakkola bound ; Multiple Instance Learning ; Pólya-Gamma ; Variational inference</subject><ispartof>Artificial intelligence, 2024-06, Vol.331, p.104115, Article 104115</ispartof><rights>2024 The Author(s)</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c352t-707041ac70c3f7b5595b2ef0940fa4eb5f1b0e7e17324979dd93fd83687d185e3</citedby><cites>FETCH-LOGICAL-c352t-707041ac70c3f7b5595b2ef0940fa4eb5f1b0e7e17324979dd93fd83687d185e3</cites><orcidid>0000-0001-6980-9746 ; 0000-0003-4694-8588 ; 0000-0003-2793-0083 ; 0000-0002-9040-3641</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,780,784,27924,27925</link.rule.ids></links><search><creatorcontrib>Castro-Macías, Francisco M.</creatorcontrib><creatorcontrib>Morales-Álvarez, Pablo</creatorcontrib><creatorcontrib>Wu, Yunan</creatorcontrib><creatorcontrib>Molina, Rafael</creatorcontrib><creatorcontrib>Katsaggelos, Aggelos K.</creatorcontrib><title>Hyperbolic Secant representation of the logistic function: Application to probabilistic Multiple Instance Learning for CT intracranial hemorrhage detection</title><title>Artificial intelligence</title><description>Multiple Instance Learning (MIL) is a weakly supervised paradigm that has been successfully applied to many different scientific areas and is particularly well suited to medical imaging. Probabilistic MIL methods, and more specifically Gaussian Processes (GPs), have achieved excellent results due to their high expressiveness and uncertainty quantification capabilities. One of the most successful GP-based MIL methods, VGPMIL, resorts to a variational bound to handle the intractability of the logistic function. Here, we formulate VGPMIL using Pólya-Gamma random variables. This approach yields the same variational posterior approximations as the original VGPMIL, which is a consequence of the two representations that the Hyperbolic Secant distribution admits. This leads us to propose a general GP-based MIL method that takes different forms by simply leveraging distributions other than the Hyperbolic Secant one. Using the Gamma distribution we arrive at a new approach that obtains competitive or superior predictive performance and efficiency. This is validated in a comprehensive experimental study including one synthetic MIL dataset, two well-known MIL benchmarks, and a real-world medical problem. We expect that this work provides useful ideas beyond MIL that can foster further research in the field.</description><subject>Gaussian processes</subject><subject>Hyperbolic Secant distribution</subject><subject>Intracranial hemorrhage detection</subject><subject>Jaakkola bound</subject><subject>Multiple Instance Learning</subject><subject>Pólya-Gamma</subject><subject>Variational inference</subject><issn>0004-3702</issn><issn>1872-7921</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><recordid>eNp9UMtOwzAQtBBIlMIfcPAPpNh51A0HpKoCWqmIA-VsOc66dZXa0dpF6rfws7iEM6fV7s7szgwh95xNOOPTh_1EYbQuTnKWl2lUcl5dkBGfiTwTdc4vyYgxVmaFYPk1uQlhn9qirvmIfC9PPWDjO6vpB2jlIkXoEQK4qKL1jnpD4w5o57c2xIQyR6fPi0c67_tEG1DR0x59oxrbDbC3Yxdt3wFduRCV00DXoNBZt6XGI11saBKMSqNyVnV0BwePuFNboC1E-P1wS66M6gLc_dUx-Xx53iyW2fr9dbWYrzNdVHnMBBPJsdKC6cKIpqrqqsnBsLpkRpXQVIY3DARwUeRlLeq2rQvTzorpTLR8VkExJuVwV6MPAcHIHu1B4UlyJs8By70cApbngOUQcKI9DTRI2r4soAzaQnLaWkwGZOvt_wd-AMulioY</recordid><startdate>202406</startdate><enddate>202406</enddate><creator>Castro-Macías, Francisco M.</creator><creator>Morales-Álvarez, Pablo</creator><creator>Wu, Yunan</creator><creator>Molina, Rafael</creator><creator>Katsaggelos, Aggelos K.</creator><general>Elsevier B.V</general><scope>6I.</scope><scope>AAFTH</scope><scope>AAYXX</scope><scope>CITATION</scope><orcidid>https://orcid.org/0000-0001-6980-9746</orcidid><orcidid>https://orcid.org/0000-0003-4694-8588</orcidid><orcidid>https://orcid.org/0000-0003-2793-0083</orcidid><orcidid>https://orcid.org/0000-0002-9040-3641</orcidid></search><sort><creationdate>202406</creationdate><title>Hyperbolic Secant representation of the logistic function: Application to probabilistic Multiple Instance Learning for CT intracranial hemorrhage detection</title><author>Castro-Macías, Francisco M. ; Morales-Álvarez, Pablo ; Wu, Yunan ; Molina, Rafael ; Katsaggelos, Aggelos K.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c352t-707041ac70c3f7b5595b2ef0940fa4eb5f1b0e7e17324979dd93fd83687d185e3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Gaussian processes</topic><topic>Hyperbolic Secant distribution</topic><topic>Intracranial hemorrhage detection</topic><topic>Jaakkola bound</topic><topic>Multiple Instance Learning</topic><topic>Pólya-Gamma</topic><topic>Variational inference</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Castro-Macías, Francisco M.</creatorcontrib><creatorcontrib>Morales-Álvarez, Pablo</creatorcontrib><creatorcontrib>Wu, Yunan</creatorcontrib><creatorcontrib>Molina, Rafael</creatorcontrib><creatorcontrib>Katsaggelos, Aggelos K.</creatorcontrib><collection>ScienceDirect Open Access Titles</collection><collection>Elsevier:ScienceDirect:Open Access</collection><collection>CrossRef</collection><jtitle>Artificial intelligence</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Castro-Macías, Francisco M.</au><au>Morales-Álvarez, Pablo</au><au>Wu, Yunan</au><au>Molina, Rafael</au><au>Katsaggelos, Aggelos K.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Hyperbolic Secant representation of the logistic function: Application to probabilistic Multiple Instance Learning for CT intracranial hemorrhage detection</atitle><jtitle>Artificial intelligence</jtitle><date>2024-06</date><risdate>2024</risdate><volume>331</volume><spage>104115</spage><pages>104115-</pages><artnum>104115</artnum><issn>0004-3702</issn><eissn>1872-7921</eissn><abstract>Multiple Instance Learning (MIL) is a weakly supervised paradigm that has been successfully applied to many different scientific areas and is particularly well suited to medical imaging. Probabilistic MIL methods, and more specifically Gaussian Processes (GPs), have achieved excellent results due to their high expressiveness and uncertainty quantification capabilities. One of the most successful GP-based MIL methods, VGPMIL, resorts to a variational bound to handle the intractability of the logistic function. Here, we formulate VGPMIL using Pólya-Gamma random variables. This approach yields the same variational posterior approximations as the original VGPMIL, which is a consequence of the two representations that the Hyperbolic Secant distribution admits. This leads us to propose a general GP-based MIL method that takes different forms by simply leveraging distributions other than the Hyperbolic Secant one. Using the Gamma distribution we arrive at a new approach that obtains competitive or superior predictive performance and efficiency. This is validated in a comprehensive experimental study including one synthetic MIL dataset, two well-known MIL benchmarks, and a real-world medical problem. We expect that this work provides useful ideas beyond MIL that can foster further research in the field.</abstract><pub>Elsevier B.V</pub><doi>10.1016/j.artint.2024.104115</doi><orcidid>https://orcid.org/0000-0001-6980-9746</orcidid><orcidid>https://orcid.org/0000-0003-4694-8588</orcidid><orcidid>https://orcid.org/0000-0003-2793-0083</orcidid><orcidid>https://orcid.org/0000-0002-9040-3641</orcidid><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | ISSN: 0004-3702 |
ispartof | Artificial intelligence, 2024-06, Vol.331, p.104115, Article 104115 |
issn | 0004-3702 1872-7921 |
language | eng |
recordid | cdi_crossref_primary_10_1016_j_artint_2024_104115 |
source | Elsevier:Jisc Collections:Elsevier Read and Publish Agreement 2022-2024:Freedom Collection (Reading list) |
subjects | Gaussian processes Hyperbolic Secant distribution Intracranial hemorrhage detection Jaakkola bound Multiple Instance Learning Pólya-Gamma Variational inference |
title | Hyperbolic Secant representation of the logistic function: Application to probabilistic Multiple Instance Learning for CT intracranial hemorrhage detection |
url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-25T02%3A37%3A06IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-elsevier_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Hyperbolic%20Secant%20representation%20of%20the%20logistic%20function:%20Application%20to%20probabilistic%20Multiple%20Instance%20Learning%20for%20CT%20intracranial%20hemorrhage%20detection&rft.jtitle=Artificial%20intelligence&rft.au=Castro-Mac%C3%ADas,%20Francisco%20M.&rft.date=2024-06&rft.volume=331&rft.spage=104115&rft.pages=104115-&rft.artnum=104115&rft.issn=0004-3702&rft.eissn=1872-7921&rft_id=info:doi/10.1016/j.artint.2024.104115&rft_dat=%3Celsevier_cross%3ES0004370224000511%3C/elsevier_cross%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c352t-707041ac70c3f7b5595b2ef0940fa4eb5f1b0e7e17324979dd93fd83687d185e3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |