Normalized total loss on the Facebook dataset (DCOR with and without RLC).

<p>To enable a fair visual comparison of training dynamics, each curve is normalized as in <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0335135#pone.0335135.e243" target="_blank">Eq (29)</a> by dividing by its epoch-1 value and EMA-smoot...

Celý popis

Uloženo v:
Podrobná bibliografie
Hlavní autor: Hossein Rafieizadeh (22676722) (author)
Další autoři: Hadi Zare (20073000) (author), Mohsen Ghassemi Parsa (22676725) (author), Hocine Cherifi (8177628) (author)
Vydáno: 2025
Témata:
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
_version_ 1849927640793743360
author Hossein Rafieizadeh (22676722)
author2 Hadi Zare (20073000)
Mohsen Ghassemi Parsa (22676725)
Hocine Cherifi (8177628)
author2_role author
author
author
author_facet Hossein Rafieizadeh (22676722)
Hadi Zare (20073000)
Mohsen Ghassemi Parsa (22676725)
Hocine Cherifi (8177628)
author_role author
dc.creator.none.fl_str_mv Hossein Rafieizadeh (22676722)
Hadi Zare (20073000)
Mohsen Ghassemi Parsa (22676725)
Hocine Cherifi (8177628)
dc.date.none.fl_str_mv 2025-11-24T18:37:58Z
dc.identifier.none.fl_str_mv 10.1371/journal.pone.0335135.g006
dc.relation.none.fl_str_mv https://figshare.com/articles/figure/Normalized_total_loss_on_the_Facebook_dataset_DCOR_with_and_without_RLC_/30698017
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Cell Biology
Science Policy
Environmental Sciences not elsewhere classified
Biological Sciences not elsewhere classified
Information Systems not elsewhere classified
intrusions across social
reconstructions across views
level contrastive learning
dual contrastive learning
across six benchmarks
div >< p
view discrepancies underutilized
augmented graph views
dcor improves auroc
view discrepancies
level contrast
dual autoencoder
augmented view
specific information
six datasets
reduces auroc
publicly available
preserves fine
physical domains
performing non
maximum gain
leaving cross
identifying threats
financial fraud
existing graph
dcor reconstructs
dcor ),
contrasts reconstructions
attributed networks
attribute patterns
dc.title.none.fl_str_mv Normalized total loss on the Facebook dataset (DCOR with and without RLC).
dc.type.none.fl_str_mv Image
Figure
info:eu-repo/semantics/publishedVersion
image
description <p>To enable a fair visual comparison of training dynamics, each curve is normalized as in <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0335135#pone.0335135.e243" target="_blank">Eq (29)</a> by dividing by its epoch-1 value and EMA-smoothed (exponential moving average) with . The EMA is computed as with . This normalization emphasizes relative convergence behavior (shape and stability) rather than raw magnitudes: with RLC, the total objective continues to decrease in late epochs, whereas without RLC it plateaus, consistent with the ablation trends in <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0335135#pone.0335135.t005" target="_blank">Table 5</a>.</p>
eu_rights_str_mv openAccess
id Manara_ac00d616de33d0d53c63092392b46c74
identifier_str_mv 10.1371/journal.pone.0335135.g006
network_acronym_str Manara
network_name_str ManaraRepo
oai_identifier_str oai:figshare.com:article/30698017
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling Normalized total loss on the Facebook dataset (DCOR with and without RLC).Hossein Rafieizadeh (22676722)Hadi Zare (20073000)Mohsen Ghassemi Parsa (22676725)Hocine Cherifi (8177628)Cell BiologyScience PolicyEnvironmental Sciences not elsewhere classifiedBiological Sciences not elsewhere classifiedInformation Systems not elsewhere classifiedintrusions across socialreconstructions across viewslevel contrastive learningdual contrastive learningacross six benchmarksdiv >< pview discrepancies underutilizedaugmented graph viewsdcor improves aurocview discrepancieslevel contrastdual autoencoderaugmented viewspecific informationsix datasetsreduces aurocpublicly availablepreserves finephysical domainsperforming nonmaximum gainleaving crossidentifying threatsfinancial fraudexisting graphdcor reconstructsdcor ),contrasts reconstructionsattributed networksattribute patterns<p>To enable a fair visual comparison of training dynamics, each curve is normalized as in <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0335135#pone.0335135.e243" target="_blank">Eq (29)</a> by dividing by its epoch-1 value and EMA-smoothed (exponential moving average) with . The EMA is computed as with . This normalization emphasizes relative convergence behavior (shape and stability) rather than raw magnitudes: with RLC, the total objective continues to decrease in late epochs, whereas without RLC it plateaus, consistent with the ablation trends in <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0335135#pone.0335135.t005" target="_blank">Table 5</a>.</p>2025-11-24T18:37:58ZImageFigureinfo:eu-repo/semantics/publishedVersionimage10.1371/journal.pone.0335135.g006https://figshare.com/articles/figure/Normalized_total_loss_on_the_Facebook_dataset_DCOR_with_and_without_RLC_/30698017CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/306980172025-11-24T18:37:58Z
spellingShingle Normalized total loss on the Facebook dataset (DCOR with and without RLC).
Hossein Rafieizadeh (22676722)
Cell Biology
Science Policy
Environmental Sciences not elsewhere classified
Biological Sciences not elsewhere classified
Information Systems not elsewhere classified
intrusions across social
reconstructions across views
level contrastive learning
dual contrastive learning
across six benchmarks
div >< p
view discrepancies underutilized
augmented graph views
dcor improves auroc
view discrepancies
level contrast
dual autoencoder
augmented view
specific information
six datasets
reduces auroc
publicly available
preserves fine
physical domains
performing non
maximum gain
leaving cross
identifying threats
financial fraud
existing graph
dcor reconstructs
dcor ),
contrasts reconstructions
attributed networks
attribute patterns
status_str publishedVersion
title Normalized total loss on the Facebook dataset (DCOR with and without RLC).
title_full Normalized total loss on the Facebook dataset (DCOR with and without RLC).
title_fullStr Normalized total loss on the Facebook dataset (DCOR with and without RLC).
title_full_unstemmed Normalized total loss on the Facebook dataset (DCOR with and without RLC).
title_short Normalized total loss on the Facebook dataset (DCOR with and without RLC).
title_sort Normalized total loss on the Facebook dataset (DCOR with and without RLC).
topic Cell Biology
Science Policy
Environmental Sciences not elsewhere classified
Biological Sciences not elsewhere classified
Information Systems not elsewhere classified
intrusions across social
reconstructions across views
level contrastive learning
dual contrastive learning
across six benchmarks
div >< p
view discrepancies underutilized
augmented graph views
dcor improves auroc
view discrepancies
level contrast
dual autoencoder
augmented view
specific information
six datasets
reduces auroc
publicly available
preserves fine
physical domains
performing non
maximum gain
leaving cross
identifying threats
financial fraud
existing graph
dcor reconstructs
dcor ),
contrasts reconstructions
attributed networks
attribute patterns