Deep Temporal and Structural Embeddings for Robust Unsupervised Anomaly Detection in Dynamic Graphs

<p dir="ltr">Detecting anomalies in dynamic graphs is a complex yet essential task, as existing methods often fail to capture long-term dependencies required for identifying irregularities in evolving networks. We introduce Temporal Structural Graph Anomaly Detection (T-StructGAD), a...

وصف كامل

محفوظ في:
التفاصيل البيبلوغرافية
المؤلف الرئيسي: Samir Abdaljalil (11513178) (author)
مؤلفون آخرون: Hasan Kurban (13144983) (author), Rachad Atat (16864194) (author), Erchin Serpedin (3706543) (author), Khalid Qaraqe (16896504) (author)
منشور في: 2025
الموضوعات:
الوسوم: إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
_version_ 1864513531589689344
author Samir Abdaljalil (11513178)
author2 Hasan Kurban (13144983)
Rachad Atat (16864194)
Erchin Serpedin (3706543)
Khalid Qaraqe (16896504)
author2_role author
author
author
author
author_facet Samir Abdaljalil (11513178)
Hasan Kurban (13144983)
Rachad Atat (16864194)
Erchin Serpedin (3706543)
Khalid Qaraqe (16896504)
author_role author
dc.creator.none.fl_str_mv Samir Abdaljalil (11513178)
Hasan Kurban (13144983)
Rachad Atat (16864194)
Erchin Serpedin (3706543)
Khalid Qaraqe (16896504)
dc.date.none.fl_str_mv 2025-07-25T12:00:00Z
dc.identifier.none.fl_str_mv 10.1109/ojcs.2025.3584942
dc.relation.none.fl_str_mv https://figshare.com/articles/journal_contribution/Deep_Temporal_and_Structural_Embeddings_for_Robust_Unsupervised_Anomaly_Detection_in_Dynamic_Graphs/30860093
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Information and computing sciences
Artificial intelligence
Computer vision and multimedia computation
Machine learning
Anomaly detection
deep learning
graph neural networks
dynamic graphs
node embedding
spatial-temporal dependencies
Long short term memory
Logic gates
Vectors
Deep learning
Autoencoders
Image edge detection
Gated recurrent units
Feature extraction
Computer architecture
dc.title.none.fl_str_mv Deep Temporal and Structural Embeddings for Robust Unsupervised Anomaly Detection in Dynamic Graphs
dc.type.none.fl_str_mv Text
Journal contribution
info:eu-repo/semantics/publishedVersion
text
contribution to journal
description <p dir="ltr">Detecting anomalies in dynamic graphs is a complex yet essential task, as existing methods often fail to capture long-term dependencies required for identifying irregularities in evolving networks. We introduce Temporal Structural Graph Anomaly Detection (T-StructGAD), an unsupervised framework that leverages Graph Convolutional Gated Recurrent Units (GConvGRUs) and Long Short-Term Memory networks (LSTMs) to jointly model both structural and temporal dynamics in graph node embeddings. Anomalies are detected using reconstruction errors generated by an AutoEncoder, enabling the framework to robustly uncover deviations across time. Our method successfully captures temporal patterns, making it robust against subtle anomalies and structural changes. Comprehensive evaluations on four real-world datasets demonstrate that T-StructGAD consistently outperforms 12 state-of-the-art unsupervised anomaly detection models, showcasing its superior ability to detect complex anomalies in evolving graphs. This work advances anomaly detection in dynamic graphs by integrating deep learning techniques to address structural and temporal irregularities in a more effective manner.</p><h2 dir="ltr">Other Information</h2><p dir="ltr">Published in: IEEE Open Journal of the Computer Society<br>License: <a href="https://creativecommons.org/licenses/by/4.0/deed.en" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1109/ojcs.2025.3584942" target="_blank">https://dx.doi.org/10.1109/ojcs.2025.3584942</a></p>
eu_rights_str_mv openAccess
id Manara2_8efdd0c8fa389428896d93b17c2f2341
identifier_str_mv 10.1109/ojcs.2025.3584942
network_acronym_str Manara2
network_name_str Manara2
oai_identifier_str oai:figshare.com:article/30860093
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling Deep Temporal and Structural Embeddings for Robust Unsupervised Anomaly Detection in Dynamic GraphsSamir Abdaljalil (11513178)Hasan Kurban (13144983)Rachad Atat (16864194)Erchin Serpedin (3706543)Khalid Qaraqe (16896504)Information and computing sciencesArtificial intelligenceComputer vision and multimedia computationMachine learningAnomaly detectiondeep learninggraph neural networksdynamic graphsnode embeddingspatial-temporal dependenciesLong short term memoryLogic gatesVectorsDeep learningAutoencodersImage edge detectionGated recurrent unitsFeature extractionComputer architecture<p dir="ltr">Detecting anomalies in dynamic graphs is a complex yet essential task, as existing methods often fail to capture long-term dependencies required for identifying irregularities in evolving networks. We introduce Temporal Structural Graph Anomaly Detection (T-StructGAD), an unsupervised framework that leverages Graph Convolutional Gated Recurrent Units (GConvGRUs) and Long Short-Term Memory networks (LSTMs) to jointly model both structural and temporal dynamics in graph node embeddings. Anomalies are detected using reconstruction errors generated by an AutoEncoder, enabling the framework to robustly uncover deviations across time. Our method successfully captures temporal patterns, making it robust against subtle anomalies and structural changes. Comprehensive evaluations on four real-world datasets demonstrate that T-StructGAD consistently outperforms 12 state-of-the-art unsupervised anomaly detection models, showcasing its superior ability to detect complex anomalies in evolving graphs. This work advances anomaly detection in dynamic graphs by integrating deep learning techniques to address structural and temporal irregularities in a more effective manner.</p><h2 dir="ltr">Other Information</h2><p dir="ltr">Published in: IEEE Open Journal of the Computer Society<br>License: <a href="https://creativecommons.org/licenses/by/4.0/deed.en" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1109/ojcs.2025.3584942" target="_blank">https://dx.doi.org/10.1109/ojcs.2025.3584942</a></p>2025-07-25T12:00:00ZTextJournal contributioninfo:eu-repo/semantics/publishedVersiontextcontribution to journal10.1109/ojcs.2025.3584942https://figshare.com/articles/journal_contribution/Deep_Temporal_and_Structural_Embeddings_for_Robust_Unsupervised_Anomaly_Detection_in_Dynamic_Graphs/30860093CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/308600932025-07-25T12:00:00Z
spellingShingle Deep Temporal and Structural Embeddings for Robust Unsupervised Anomaly Detection in Dynamic Graphs
Samir Abdaljalil (11513178)
Information and computing sciences
Artificial intelligence
Computer vision and multimedia computation
Machine learning
Anomaly detection
deep learning
graph neural networks
dynamic graphs
node embedding
spatial-temporal dependencies
Long short term memory
Logic gates
Vectors
Deep learning
Autoencoders
Image edge detection
Gated recurrent units
Feature extraction
Computer architecture
status_str publishedVersion
title Deep Temporal and Structural Embeddings for Robust Unsupervised Anomaly Detection in Dynamic Graphs
title_full Deep Temporal and Structural Embeddings for Robust Unsupervised Anomaly Detection in Dynamic Graphs
title_fullStr Deep Temporal and Structural Embeddings for Robust Unsupervised Anomaly Detection in Dynamic Graphs
title_full_unstemmed Deep Temporal and Structural Embeddings for Robust Unsupervised Anomaly Detection in Dynamic Graphs
title_short Deep Temporal and Structural Embeddings for Robust Unsupervised Anomaly Detection in Dynamic Graphs
title_sort Deep Temporal and Structural Embeddings for Robust Unsupervised Anomaly Detection in Dynamic Graphs
topic Information and computing sciences
Artificial intelligence
Computer vision and multimedia computation
Machine learning
Anomaly detection
deep learning
graph neural networks
dynamic graphs
node embedding
spatial-temporal dependencies
Long short term memory
Logic gates
Vectors
Deep learning
Autoencoders
Image edge detection
Gated recurrent units
Feature extraction
Computer architecture