Region-Based Active Learning with Hierarchical and Adaptive Region Construction
Learning of classification models in practice often relies on human annotation effort in which humans assign class labels to data instances. As this process can be very time-consuming and costly, finding effective ways to reduce the annotation cost becomes critical for building such models. To solve...
Uložené v:
| Vydané v: | Proceedings of the ... SIAM International Conference on Data Mining Ročník 2019; s. 441 |
|---|---|
| Hlavní autori: | , |
| Médium: | Journal Article |
| Jazyk: | English |
| Vydavateľské údaje: |
United States
01.05.2019
|
| ISSN: | 2167-0102 |
| On-line prístup: | Zistit podrobnosti o prístupe |
| Tagy: |
Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
|
| Abstract | Learning of classification models in practice often relies on human annotation effort in which humans assign class labels to data instances. As this process can be very time-consuming and costly, finding effective ways to reduce the annotation cost becomes critical for building such models. To solve this problem, instead of soliciting instance-based annotation we explore
-based annotation as the human feedback. A region is defined as a hyper-cubic subspace of the input space
and it covers a subpopulation of data instances that fall into this region. Each region is labeled with a number in [0,1] (in binary classification setting), representing a human estimate of the positive (or negative) class proportion in the subpopulation. To quickly discover pure regions (in terms of class proportion) in the data, we have developed a novel active learning framework that constructs regions in a
and
way.
means that regions are incrementally built into a hierarchical tree, which is done by repeatedly splitting the input space.
means that our framework can adaptively choose the best heuristic for each of the region splits. Through experiments on numerous datasets we demonstrate that our framework can identify pure regions in very few region queries. Thus our approach is shown to be effective in learning classification models from very limited human feedback. |
|---|---|
| AbstractList | Learning of classification models in practice often relies on human annotation effort in which humans assign class labels to data instances. As this process can be very time-consuming and costly, finding effective ways to reduce the annotation cost becomes critical for building such models. To solve this problem, instead of soliciting instance-based annotation we explore
-based annotation as the human feedback. A region is defined as a hyper-cubic subspace of the input space
and it covers a subpopulation of data instances that fall into this region. Each region is labeled with a number in [0,1] (in binary classification setting), representing a human estimate of the positive (or negative) class proportion in the subpopulation. To quickly discover pure regions (in terms of class proportion) in the data, we have developed a novel active learning framework that constructs regions in a
and
way.
means that regions are incrementally built into a hierarchical tree, which is done by repeatedly splitting the input space.
means that our framework can adaptively choose the best heuristic for each of the region splits. Through experiments on numerous datasets we demonstrate that our framework can identify pure regions in very few region queries. Thus our approach is shown to be effective in learning classification models from very limited human feedback. Learning of classification models in practice often relies on human annotation effort in which humans assign class labels to data instances. As this process can be very time-consuming and costly, finding effective ways to reduce the annotation cost becomes critical for building such models. To solve this problem, instead of soliciting instance-based annotation we explore region-based annotation as the human feedback. A region is defined as a hyper-cubic subspace of the input space X and it covers a subpopulation of data instances that fall into this region. Each region is labeled with a number in [0,1] (in binary classification setting), representing a human estimate of the positive (or negative) class proportion in the subpopulation. To quickly discover pure regions (in terms of class proportion) in the data, we have developed a novel active learning framework that constructs regions in a hierarchical and adaptive way. Hierarchical means that regions are incrementally built into a hierarchical tree, which is done by repeatedly splitting the input space. Adaptive means that our framework can adaptively choose the best heuristic for each of the region splits. Through experiments on numerous datasets we demonstrate that our framework can identify pure regions in very few region queries. Thus our approach is shown to be effective in learning classification models from very limited human feedback.Learning of classification models in practice often relies on human annotation effort in which humans assign class labels to data instances. As this process can be very time-consuming and costly, finding effective ways to reduce the annotation cost becomes critical for building such models. To solve this problem, instead of soliciting instance-based annotation we explore region-based annotation as the human feedback. A region is defined as a hyper-cubic subspace of the input space X and it covers a subpopulation of data instances that fall into this region. Each region is labeled with a number in [0,1] (in binary classification setting), representing a human estimate of the positive (or negative) class proportion in the subpopulation. To quickly discover pure regions (in terms of class proportion) in the data, we have developed a novel active learning framework that constructs regions in a hierarchical and adaptive way. Hierarchical means that regions are incrementally built into a hierarchical tree, which is done by repeatedly splitting the input space. Adaptive means that our framework can adaptively choose the best heuristic for each of the region splits. Through experiments on numerous datasets we demonstrate that our framework can identify pure regions in very few region queries. Thus our approach is shown to be effective in learning classification models from very limited human feedback. |
| Author | Hauskrecht, Milos Luo, Zhipeng |
| Author_xml | – sequence: 1 givenname: Zhipeng surname: Luo fullname: Luo, Zhipeng organization: Department of Computer Science, University of Pittsburgh, PA, USA – sequence: 2 givenname: Milos surname: Hauskrecht fullname: Hauskrecht, Milos organization: Department of Computer Science, University of Pittsburgh, PA, USA |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/31929950$$D View this record in MEDLINE/PubMed |
| BookMark | eNo1kLFOwzAURT0U0VL6AwwoI0uKnx3H8VgqoEiVKiGYo1f7pTVKnRAnIP6eipbp3uGcO9wrNgpNIMZugM8BpL6HudEF5ABGq1zLueIjNhGQ65QDF2M2i_GD82NXOhP8ko0lGGGM4hO2eaWdb0L6gJFcsrC9_6JkTdgFH3bJt-_3ycpTh53de4t1guFIOWz_uJObLJsQ-244uk24ZhcV1pFm55yy96fHt-UqXW-eX5aLddpK4H1KljRkhYTKZqgKC0Rk8xzQUUWoNKotmqKqJLoKxNZokgXmDi2i45lTYsruTrtt13wOFPvy4KOlusZAzRBLIWXBteFCH9HbMzpsD-TKtvMH7H7K_xPELwBrYNQ |
| ContentType | Journal Article |
| DBID | NPM 7X8 |
| DOI | 10.1137/1.9781611975673.50 |
| DatabaseName | PubMed MEDLINE - Academic |
| DatabaseTitle | PubMed MEDLINE - Academic |
| DatabaseTitleList | PubMed MEDLINE - Academic |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | no_fulltext_linktorsrc |
| Discipline | Computer Science |
| ExternalDocumentID | 31929950 |
| Genre | Journal Article |
| GrantInformation_xml | – fundername: NIGMS NIH HHS grantid: R01 GM088224 – fundername: NLM NIH HHS grantid: R01 LM010019 |
| GroupedDBID | 3V. 7WY 7X2 7XC 88A 88I 8CJ 8FE 8FG 8FH 8FL 8G5 ABJCF ABUWG ACGOD ACIWK ACPRK ADBBV AFKRA AFRAH ALMA_UNASSIGNED_HOLDINGS ARAPS ATCPS AZQEC BBNVY BENPR BEZIV BGLVJ BHPHI BPHCQ CCPQU CZ9 D1I D1J D1K DWQXO FRNLG GNUQQ GROUPED_ABI_INFORM_COMPLETE GUQSH HCIFZ K6- K60 K6V K6~ K7- KB. KC. L6V LK5 LK8 M0C M0K M0L M0N M1Q M2O M2P M7P M7R M7S NPM P62 PATMY PDBOC PQBIZ PQBZA PQQKQ PROAC PTHSS PYCSY 7X8 |
| ID | FETCH-LOGICAL-p310t-ece714831fc4a58c1eeec661adefea57a5ba98ff3adf12b97e38a6dacaad04d52 |
| IEDL.DBID | 7X8 |
| ISICitedReferencesCount | 4 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001288498300045&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 2167-0102 |
| IngestDate | Fri Jul 11 11:06:54 EDT 2025 Thu Jan 02 23:00:21 EST 2025 |
| IsDoiOpenAccess | false |
| IsOpenAccess | true |
| IsPeerReviewed | false |
| IsScholarly | false |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-p310t-ece714831fc4a58c1eeec661adefea57a5ba98ff3adf12b97e38a6dacaad04d52 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| OpenAccessLink | https://epubs.siam.org/doi/pdf/10.1137/1.9781611975673.50 |
| PMID | 31929950 |
| PQID | 2338079027 |
| PQPubID | 23479 |
| ParticipantIDs | proquest_miscellaneous_2338079027 pubmed_primary_31929950 |
| PublicationCentury | 2000 |
| PublicationDate | 20190501 |
| PublicationDateYYYYMMDD | 2019-05-01 |
| PublicationDate_xml | – month: 5 year: 2019 text: 20190501 day: 1 |
| PublicationDecade | 2010 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States |
| PublicationTitle | Proceedings of the ... SIAM International Conference on Data Mining |
| PublicationTitleAlternate | Proc SIAM Int Conf Data Min |
| PublicationYear | 2019 |
| SSID | ssj0001057420 |
| Score | 1.7318677 |
| Snippet | Learning of classification models in practice often relies on human annotation effort in which humans assign class labels to data instances. As this process... |
| SourceID | proquest pubmed |
| SourceType | Aggregation Database Index Database |
| StartPage | 441 |
| Title | Region-Based Active Learning with Hierarchical and Adaptive Region Construction |
| URI | https://www.ncbi.nlm.nih.gov/pubmed/31929950 https://www.proquest.com/docview/2338079027 |
| Volume | 2019 |
| WOSCitedRecordID | wos001288498300045&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LSwMxEA5qPXixvq0vInjddneTbHZPUsXSg9YiKr2VSTKRgmxXW_39JvugJ0HwspdNIAyTmS-TyfcRcpVIiVFmlXNezgMOnAUqlSwQCFL7X2HJzv96L0ejdDLJxnXBbVG3VTYxsQzUZq59jbwXM0-NnrlT1HXxEXjVKH-7WktorJMWc1DGt3TJSbqqsTgwwktmxtjTe3v6tObdDJO9qOv5nhJ_jyYSyboi_B1lltlm0P7vOnfIdo0zab9yjF2yhvkeaTcaDrTe0vvk8Ql9R3Jw47KZof0y-tGadPWN-iotHc78I-VSM-WdQu5GGSjKcdVc6kU_GxraA_IyuHu-HQa1yEJQOGS3DFCjdEciFlnNQaQ6QkTtkjYYtAhCglCQpdYyMDaKVSaRpZAY0AAm5EbEh2Qjn-d4TKixEFplGAJTnMVGxUqmnKNUiQlBhx1y2Zhs6pzY30xAjvOvxXRltA45quw-LSq2jamLES5livDkD7NPyZYDNFnVkHhGWtZtYTwnm_p7OVt8XpTe4b6j8cMPf0vFLA |
| linkProvider | ProQuest |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Region-Based+Active+Learning+with+Hierarchical+and+Adaptive+Region+Construction&rft.jtitle=Proceedings+of+the+...+SIAM+International+Conference+on+Data+Mining&rft.au=Luo%2C+Zhipeng&rft.au=Hauskrecht%2C+Milos&rft.date=2019-05-01&rft.issn=2167-0102&rft.volume=2019&rft.spage=441&rft_id=info:doi/10.1137%2F1.9781611975673.50&rft_id=info%3Apmid%2F31929950&rft_id=info%3Apmid%2F31929950&rft.externalDocID=31929950 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2167-0102&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2167-0102&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2167-0102&client=summon |