Deep Clustering by Semantic Contrastive Learning
Whilst contrastive learning has recently brought notable benefits to deep clustering of unlabelled images by learning sample-specific discriminative visual features, its potential for explicitly inferring class decision boundaries is less well understood. This is because its instance discrimination...
Gespeichert in:
Veröffentlicht in: | arXiv.org 2022-11 |
---|---|
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 | Huang, Jiabo Gong, Shaogang |
description | Whilst contrastive learning has recently brought notable benefits to deep clustering of unlabelled images by learning sample-specific discriminative visual features, its potential for explicitly inferring class decision boundaries is less well understood. This is because its instance discrimination strategy is not class sensitive, therefore, the clusters derived on the resulting sample-specific feature space are not optimised for corresponding to meaningful class decision boundaries. In this work, we solve this problem by introducing Semantic Contrastive Learning (SCL). SCL imposes explicitly distance-based cluster structures on unlabelled training data by formulating a semantic (cluster-aware) contrastive learning objective. Moreover, we introduce a clustering consistency condition to be satisfied jointly by both instance visual similarities and cluster decision boundaries, and concurrently optimising both to reason about the hypotheses of semantic ground-truth classes (unknown/unlabelled) on-the-fly by their consensus. This semantic contrastive learning approach to discovering unknown class decision boundaries has considerable advantages to unsupervised learning of object recognition tasks. Extensive experiments show that SCL outperforms state-of-the-art contrastive learning and deep clustering methods on six object recognition benchmarks, especially on the more challenging finer-grained and larger datasets. |
format | Article |
fullrecord | <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2497372851</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2497372851</sourcerecordid><originalsourceid>FETCH-proquest_journals_24973728513</originalsourceid><addsrcrecordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mQwcElNLVBwziktLkktysxLV0iqVAhOzU3MK8lMVnDOzyspSiwuySxLVfBJTSzKAyrgYWBNS8wpTuWF0twMym6uIc4eugVF-YWlqcUl8Vn5pUV5QKl4IxNLc2NzIwtTQ2PiVAEAZE4zJA</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2497372851</pqid></control><display><type>article</type><title>Deep Clustering by Semantic Contrastive Learning</title><source>Free E- Journals</source><creator>Huang, Jiabo ; Gong, Shaogang</creator><creatorcontrib>Huang, Jiabo ; Gong, Shaogang</creatorcontrib><description>Whilst contrastive learning has recently brought notable benefits to deep clustering of unlabelled images by learning sample-specific discriminative visual features, its potential for explicitly inferring class decision boundaries is less well understood. This is because its instance discrimination strategy is not class sensitive, therefore, the clusters derived on the resulting sample-specific feature space are not optimised for corresponding to meaningful class decision boundaries. In this work, we solve this problem by introducing Semantic Contrastive Learning (SCL). SCL imposes explicitly distance-based cluster structures on unlabelled training data by formulating a semantic (cluster-aware) contrastive learning objective. Moreover, we introduce a clustering consistency condition to be satisfied jointly by both instance visual similarities and cluster decision boundaries, and concurrently optimising both to reason about the hypotheses of semantic ground-truth classes (unknown/unlabelled) on-the-fly by their consensus. This semantic contrastive learning approach to discovering unknown class decision boundaries has considerable advantages to unsupervised learning of object recognition tasks. Extensive experiments show that SCL outperforms state-of-the-art contrastive learning and deep clustering methods on six object recognition benchmarks, especially on the more challenging finer-grained and larger datasets.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Boundaries ; Clustering ; Object recognition ; Representations ; Semantics ; Unsupervised learning</subject><ispartof>arXiv.org, 2022-11</ispartof><rights>2022. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (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>776,780</link.rule.ids></links><search><creatorcontrib>Huang, Jiabo</creatorcontrib><creatorcontrib>Gong, Shaogang</creatorcontrib><title>Deep Clustering by Semantic Contrastive Learning</title><title>arXiv.org</title><description>Whilst contrastive learning has recently brought notable benefits to deep clustering of unlabelled images by learning sample-specific discriminative visual features, its potential for explicitly inferring class decision boundaries is less well understood. This is because its instance discrimination strategy is not class sensitive, therefore, the clusters derived on the resulting sample-specific feature space are not optimised for corresponding to meaningful class decision boundaries. In this work, we solve this problem by introducing Semantic Contrastive Learning (SCL). SCL imposes explicitly distance-based cluster structures on unlabelled training data by formulating a semantic (cluster-aware) contrastive learning objective. Moreover, we introduce a clustering consistency condition to be satisfied jointly by both instance visual similarities and cluster decision boundaries, and concurrently optimising both to reason about the hypotheses of semantic ground-truth classes (unknown/unlabelled) on-the-fly by their consensus. This semantic contrastive learning approach to discovering unknown class decision boundaries has considerable advantages to unsupervised learning of object recognition tasks. Extensive experiments show that SCL outperforms state-of-the-art contrastive learning and deep clustering methods on six object recognition benchmarks, especially on the more challenging finer-grained and larger datasets.</description><subject>Boundaries</subject><subject>Clustering</subject><subject>Object recognition</subject><subject>Representations</subject><subject>Semantics</subject><subject>Unsupervised learning</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>BENPR</sourceid><recordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mQwcElNLVBwziktLkktysxLV0iqVAhOzU3MK8lMVnDOzyspSiwuySxLVfBJTSzKAyrgYWBNS8wpTuWF0twMym6uIc4eugVF-YWlqcUl8Vn5pUV5QKl4IxNLc2NzIwtTQ2PiVAEAZE4zJA</recordid><startdate>20221120</startdate><enddate>20221120</enddate><creator>Huang, Jiabo</creator><creator>Gong, Shaogang</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>20221120</creationdate><title>Deep Clustering by Semantic Contrastive Learning</title><author>Huang, Jiabo ; Gong, Shaogang</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_24973728513</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Boundaries</topic><topic>Clustering</topic><topic>Object recognition</topic><topic>Representations</topic><topic>Semantics</topic><topic>Unsupervised learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Huang, Jiabo</creatorcontrib><creatorcontrib>Gong, Shaogang</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>Publicly Available Content Database</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>Huang, Jiabo</au><au>Gong, Shaogang</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Deep Clustering by Semantic Contrastive Learning</atitle><jtitle>arXiv.org</jtitle><date>2022-11-20</date><risdate>2022</risdate><eissn>2331-8422</eissn><abstract>Whilst contrastive learning has recently brought notable benefits to deep clustering of unlabelled images by learning sample-specific discriminative visual features, its potential for explicitly inferring class decision boundaries is less well understood. This is because its instance discrimination strategy is not class sensitive, therefore, the clusters derived on the resulting sample-specific feature space are not optimised for corresponding to meaningful class decision boundaries. In this work, we solve this problem by introducing Semantic Contrastive Learning (SCL). SCL imposes explicitly distance-based cluster structures on unlabelled training data by formulating a semantic (cluster-aware) contrastive learning objective. Moreover, we introduce a clustering consistency condition to be satisfied jointly by both instance visual similarities and cluster decision boundaries, and concurrently optimising both to reason about the hypotheses of semantic ground-truth classes (unknown/unlabelled) on-the-fly by their consensus. This semantic contrastive learning approach to discovering unknown class decision boundaries has considerable advantages to unsupervised learning of object recognition tasks. Extensive experiments show that SCL outperforms state-of-the-art contrastive learning and deep clustering methods on six object recognition benchmarks, especially on the more challenging finer-grained and larger datasets.</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, 2022-11 |
issn | 2331-8422 |
language | eng |
recordid | cdi_proquest_journals_2497372851 |
source | Free E- Journals |
subjects | Boundaries Clustering Object recognition Representations Semantics Unsupervised learning |
title | Deep Clustering by Semantic Contrastive Learning |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-05T14%3A28%3A59IST&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=Deep%20Clustering%20by%20Semantic%20Contrastive%20Learning&rft.jtitle=arXiv.org&rft.au=Huang,%20Jiabo&rft.date=2022-11-20&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2497372851%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2497372851&rft_id=info:pmid/&rfr_iscdi=true |