Towards More Explainable Nonlinear Dimensionality Reduction: A Feature-Driven Interaction Approach

Nonlinear dimensionality reduction (NDR) techniques are widely used to visualize high-dimensional data. However, they often lack explainability, making it challenging for analysts to relate patterns in projections to original high-dimensional features. Existing interactive methods typically separate...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on visualization and computer graphics Jg. PP; S. 1 - 16
Hauptverfasser: Cho, Aeri, Jeon, Hyeon, Choe, Kiroong, Park, Seokhyeon, Seo, Jinwook
Format: Journal Article
Sprache:Englisch
Veröffentlicht: United States IEEE 16.10.2025
Schlagworte:
ISSN:1077-2626, 1941-0506, 1941-0506
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract Nonlinear dimensionality reduction (NDR) techniques are widely used to visualize high-dimensional data. However, they often lack explainability, making it challenging for analysts to relate patterns in projections to original high-dimensional features. Existing interactive methods typically separate user interactions from the feature space, treating them primarily as post-hoc explanations rather than integrating them into the exploration process. This separation limits insight generation by restricting users' understanding of how features dynamically influence projections. To address this limitation, we propose a bidirectional interaction method that directly bridges the feature space and the projections. By allowing users to adjust feature weights, our approach enables intuitive exploration of how different features shape the embedding. We also define visual semantics to quantify projection changes, enabling structured pattern discovery through automated query-based interaction. To ensure responsiveness despite the computational complexity of NDR, we employ a neural network to approximate the projection process, enhancing scalability while maintaining accuracy. We evaluated our approach through quantitative analysis, assessing accuracy and scalability. A user study with a comprehensive visual interface and case studies demonstrated its effectiveness in supporting hypothesis generation and exploratory tasks with real-world data. The results confirmed that our approach supports diverse analytical scenarios and enhances users' ability to explore and interpret high-dimensional data through interactive exploration grounded in the feature space.
AbstractList Nonlinear dimensionality reduction (NDR) techniques are widely used to visualize high-dimensional data. However, they often lack explainability, making it challenging for analysts to relate patterns in projections to original high-dimensional features. Existing interactive methods typically separate user interactions from the feature space, treating them primarily as post-hoc explanations rather than integrating them into the exploration process. This separation limits insight generation by restricting users' understanding of how features dynamically influence projections. To address this limitation, we propose a bidirectional interaction method that directly bridges the feature space and the projections. By allowing users to adjust feature weights, our approach enables intuitive exploration of how different features shape the embedding. We also define visual semantics to quantify projection changes, enabling structured pattern discovery through automated query-based interaction. To ensure responsiveness despite the computational complexity of NDR, we employ a neural network to approximate the projection process, enhancing scalability while maintaining accuracy. We evaluated our approach through quantitative analysis, assessing accuracy and scalability. A user study with a comprehensive visual interface and case studies demonstrated its effectiveness in supporting hypothesis generation and exploratory tasks with real-world data. The results confirmed that our approach supports diverse analytical scenarios and enhances users' ability to explore and interpret high-dimensional data through interactive exploration grounded in the feature space.Nonlinear dimensionality reduction (NDR) techniques are widely used to visualize high-dimensional data. However, they often lack explainability, making it challenging for analysts to relate patterns in projections to original high-dimensional features. Existing interactive methods typically separate user interactions from the feature space, treating them primarily as post-hoc explanations rather than integrating them into the exploration process. This separation limits insight generation by restricting users' understanding of how features dynamically influence projections. To address this limitation, we propose a bidirectional interaction method that directly bridges the feature space and the projections. By allowing users to adjust feature weights, our approach enables intuitive exploration of how different features shape the embedding. We also define visual semantics to quantify projection changes, enabling structured pattern discovery through automated query-based interaction. To ensure responsiveness despite the computational complexity of NDR, we employ a neural network to approximate the projection process, enhancing scalability while maintaining accuracy. We evaluated our approach through quantitative analysis, assessing accuracy and scalability. A user study with a comprehensive visual interface and case studies demonstrated its effectiveness in supporting hypothesis generation and exploratory tasks with real-world data. The results confirmed that our approach supports diverse analytical scenarios and enhances users' ability to explore and interpret high-dimensional data through interactive exploration grounded in the feature space.
Nonlinear dimensionality reduction (NDR) techniques are widely used to visualize high-dimensional data. However, they often lack explainability, making it challenging for analysts to relate patterns in projections to original high-dimensional features. Existing interactive methods typically separate user interactions from the feature space, treating them primarily as post-hoc explanations rather than integrating them into the exploration process. This separation limits insight generation by restricting users' understanding of how features dynamically influence projections. To address this limitation, we propose a bidirectional interaction method that directly bridges the feature space and the projections. By allowing users to adjust feature weights, our approach enables intuitive exploration of how different features shape the embedding. We also define visual semantics to quantify projection changes, enabling structured pattern discovery through automated query-based interaction. To ensure responsiveness despite the computational complexity of NDR, we employ a neural network to approximate the projection process, enhancing scalability while maintaining accuracy. We evaluated our approach through quantitative analysis, assessing accuracy and scalability. A user study with a comprehensive visual interface and case studies demonstrated its effectiveness in supporting hypothesis generation and exploratory tasks with real-world data. The results confirmed that our approach supports diverse analytical scenarios and enhances users' ability to explore and interpret high-dimensional data through interactive exploration grounded in the feature space.
Author Seo, Jinwook
Choe, Kiroong
Jeon, Hyeon
Park, Seokhyeon
Cho, Aeri
Author_xml – sequence: 1
  givenname: Aeri
  surname: Cho
  fullname: Cho, Aeri
  email: archo@hcil.snu.ac.kr
  organization: Seoul National University, Seoul, South Korea
– sequence: 2
  givenname: Hyeon
  surname: Jeon
  fullname: Jeon, Hyeon
  email: hj@hcil.snu.ac.kr
  organization: Seoul National University, Seoul, South Korea
– sequence: 3
  givenname: Kiroong
  surname: Choe
  fullname: Choe, Kiroong
  email: krchoe@hcil.snu.ac.kr
  organization: Seoul National University, Seoul, South Korea
– sequence: 4
  givenname: Seokhyeon
  surname: Park
  fullname: Park, Seokhyeon
  email: shpark@hcil.snu.ac.kr
  organization: Seoul National University, Seoul, South Korea
– sequence: 5
  givenname: Jinwook
  surname: Seo
  fullname: Seo, Jinwook
  email: jseo@snu.ac.kr
  organization: Seoul National University, Seoul, South Korea
BackLink https://www.ncbi.nlm.nih.gov/pubmed/41100239$$D View this record in MEDLINE/PubMed
BookMark eNpFkE1v2zAMhoWhw5p0-wEDikHHXpyKlCzbuwVJ0xbINmDIdjVkm0Y1OHIm2f3491WatD2RBJ-XBJ4pO3G9I8a-gpgBiOJy83dxPUOB6UxqRAD1gU2gUJCIVOiT2IssS1CjPmXTEP4JAUrlxSd2qmJcoCwmrNr0D8Y3gf_oPfGrx11nrDNVR_xn7zrryHi-tFtywfbOdHZ44r-pGeshjt_5nK_IDKOnZOntPTl-6wby5mXL57ud701995l9bE0X6MuxnrE_q6vN4iZZ_7q-XczXSQ0FDImsVYvKpHmattrIRpg2ryAXsUMNadaiURIVYZHVUOVSp9BmqBujdFWIJpNn7OJwN779P1IYyq0NNXWdcdSPoZTRBGIuNEb02xEdqy015c7brfFP5auXCMABqH0fgqf2DQFR7t2Xe_fl3n15dB8z54eMJaJ3HlCkUqJ8BvWVfrY
CODEN ITVGEA
ContentType Journal Article
DBID 97E
RIA
RIE
AAYXX
CITATION
NPM
7X8
DOI 10.1109/TVCG.2025.3622114
DatabaseName IEEE Xplore (IEEE)
IEEE All-Society Periodicals Package (ASPP) 1998–Present
IEEE Electronic Library (IEL)
CrossRef
PubMed
MEDLINE - Academic
DatabaseTitle CrossRef
PubMed
MEDLINE - Academic
DatabaseTitleList MEDLINE - Academic
PubMed

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: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
– sequence: 3
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISSN 1941-0506
EndPage 16
ExternalDocumentID 41100239
10_1109_TVCG_2025_3622114
11205332
Genre orig-research
Journal Article
GroupedDBID ---
-~X
.DC
0R~
29I
4.4
53G
5GY
6IK
97E
AAJGR
AASAJ
AAWTH
ABAZT
ABQJQ
ABVLG
ACGFO
ACIWK
AENEX
AGQYO
AHBIQ
AKJIK
AKQYR
ALMA_UNASSIGNED_HOLDINGS
ATWAV
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CS3
DU5
EBS
EJD
F5P
HZ~
IEDLZ
IFIPE
IPLJI
JAVBF
LAI
M43
O9-
OCL
P2P
PQQKQ
RIA
RIE
RNS
TN5
5VS
AAYXX
AETIX
AGSQL
AI.
AIBXA
ALLEH
CITATION
H~9
IFJZH
RNI
RZB
VH1
NPM
7X8
ID FETCH-LOGICAL-c191t-3c4f24a5855f6a3d0af8b1803d026157f2a4324e297c1b83651f726da46b90d73
IEDL.DBID RIE
ISSN 1077-2626
1941-0506
IngestDate Sat Oct 18 23:54:42 EDT 2025
Sun Oct 19 00:32:16 EDT 2025
Sat Nov 29 07:08:34 EST 2025
Sat Oct 25 03:12:11 EDT 2025
IsPeerReviewed true
IsScholarly true
Language English
License https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
https://doi.org/10.15223/policy-029
https://doi.org/10.15223/policy-037
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c191t-3c4f24a5855f6a3d0af8b1803d026157f2a4324e297c1b83651f726da46b90d73
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
PMID 41100239
PQID 3262228062
PQPubID 23479
PageCount 16
ParticipantIDs ieee_primary_11205332
proquest_miscellaneous_3262228062
pubmed_primary_41100239
crossref_primary_10_1109_TVCG_2025_3622114
PublicationCentury 2000
PublicationDate 2025-Oct-16
PublicationDateYYYYMMDD 2025-10-16
PublicationDate_xml – month: 10
  year: 2025
  text: 2025-Oct-16
  day: 16
PublicationDecade 2020
PublicationPlace United States
PublicationPlace_xml – name: United States
PublicationTitle IEEE transactions on visualization and computer graphics
PublicationTitleAbbrev TVCG
PublicationTitleAlternate IEEE Trans Vis Comput Graph
PublicationYear 2025
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0014489
Score 2.4615784
Snippet Nonlinear dimensionality reduction (NDR) techniques are widely used to visualize high-dimensional data. However, they often lack explainability, making it...
SourceID proquest
pubmed
crossref
ieee
SourceType Aggregation Database
Index Database
Publisher
StartPage 1
SubjectTerms Accuracy
Aerospace electronics
Computational modeling
Data structures
Data visualization
Dimensionality reduction
Explainability
Feature Space
High dimensional data
Interaction Techniques
Layout
Nonlinear Dimensionality Reduction
Scalability
Visualization
Title Towards More Explainable Nonlinear Dimensionality Reduction: A Feature-Driven Interaction Approach
URI https://ieeexplore.ieee.org/document/11205332
https://www.ncbi.nlm.nih.gov/pubmed/41100239
https://www.proquest.com/docview/3262228062
Volume PP
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVIEE
  databaseName: IEEE Electronic Library (IEL)
  customDbUrl:
  eissn: 1941-0506
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0014489
  issn: 1077-2626
  databaseCode: RIE
  dateStart: 19950101
  isFulltext: true
  titleUrlDefault: https://ieeexplore.ieee.org/
  providerName: IEEE
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV07T8MwELagYoCBZ4HyqIzEhOQ2cR5O2KqWwkKFUEHdIttxJJYUpQ2_nzs7LTB0YPPgOJHPOX_nu-8zIbdaKp8bJVgEHpKF2tcsDZRieZFHVg9Keu6yCTGZJLNZ-tKQ1S0Xxhhji89MD5s2l5_PdY1HZX3ABkgdBY-7LUTsyFrrlAHEGakrMBSMA0xvUpi-l_an78NHCAV51AN3DRFP-GcTsreqbAaYdqMZH_zzEw_JfoMo6cAtgSOyZcpjsvdLZ_CEqKktjl3QZxiOYtldw5miE6eUISs6Qpl_J9EBwJy-oqQrGu2eDijixLoybFShb6T2FNERIuig0SRvk7fxw3T4xJrLFZiGEG3JAh0WPJQQLURFLIPck0Wi_MSDFgRVkSi4RLE-w1OhfZUEceQXgse5DGOVerkITkmrnJfmnNBAB0b6MlSIZpTgiUHujzCJ1nlhTNIhd6vZzj6dhkZmYw8vzdA0GZoma0zTIW2c1Z-OzYR2yM3KQBn8AZjWkKWZ14sMACgeY3kx9Dlzlls_HaIiHg_Siw2jXpJdfDluRn58RVrLqjbXZEd_LT8WVReW2Szp2mX2DbmKzks
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV07T8MwELYQIAED70d5GokJyTRxnDhhqyilCFohVBBbZDuOxNKitOX3c2eHAkMHNg-OZfmc8_f57j4TcmGUDrnVksXgIZkwoWFZpDUryiJ2elAq8I9NyH4_fXvLnupidVcLY611yWf2Cpsull-MzBSvypqADbB0FDzuUiwED3y51ixoAEwj8ymGknEA6nUQMwyy5uD15g7III-vwGED5xF_jiH3rsp8iOmOms7GPye5SdZrTElbfhNskQU73CZrv5QGd4geuPTYMe3BcBQT7-qqKdr3Whmqom0U-vciHQDN6TOKuqLZrmmLIlKcVpa1K_SO1N0j-pII2qpVyXfJS-d2cNNl9fMKzABJm7DIiJILBXwhLhMVFYEqUx2mAbSAVsWy5Arl-izPpAl1GiVxWEqeFEokOgsKGe2RxeFoaA8IjUxkVaiERjyjJU8tVv9ImxpTlNamDXL5vdr5h1fRyB37CLIcTZOjafLaNA2yi6v607Fe0AY5_zZQDv8ABjbU0I6m4xwgKF5kBQn02feWm30tUBOPR9nhnFHPyEp30HvMH-_7D0dkFSeCR1OYHJPFSTW1J2TZfE7ex9Wp22xfwELQqg
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=Towards+More+Explainable+Nonlinear+Dimensionality+Reduction%3A+A+Feature-Driven+Interaction+Approach&rft.jtitle=IEEE+transactions+on+visualization+and+computer+graphics&rft.au=Cho%2C+Aeri&rft.au=Jeon%2C+Hyeon&rft.au=Choe%2C+Kiroong&rft.au=Park%2C+Seokhyeon&rft.date=2025-10-16&rft.eissn=1941-0506&rft.volume=PP&rft_id=info:doi/10.1109%2FTVCG.2025.3622114&rft_id=info%3Apmid%2F41100239&rft.externalDocID=41100239
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1077-2626&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1077-2626&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1077-2626&client=summon