Towards Automated Self-Supervised Learning for Truly Unsupervised Graph Anomaly Detection
Not on the map. Why?
Abstract (arXiv)
Self-supervised learning (SSL) is an emerging paradigm that exploits supervisory signals generated from the data itself, and many recent studies have leveraged SSL to conduct graph anomaly detection. However, we empirically found that three important factors can substantially impact detection performance across datasets: 1) the specific SSL strategy employed; 2) the tuning of the strategy's hyperparameters; and 3) the allocation of combination weights when using multiple strategies. Most SSL-based graph anomaly detection methods circumvent these issues by arbitrarily or selectively (i.e., guided by label information) choosing SSL strategies, hyperparameter settings, and combination weights. While an arbitrary choice may lead to subpar performance, using label information in an unsupervised setting is label information leakage and leads to severe overestimation of a method's performance. Leakage has been criticized as "one of the top ten data mining mistakes", yet many recent studies on SSL-based graph anomaly detection have been using label information to select hyperparameters. To mitigate this issue, we propose to use an internal evaluation strategy (with theoretical analysis) to select hyperparameters in SSL for unsupervised anomaly detection. We perform extensive experiments using 10 recent SSL-based graph anomaly detection algorithms on various benchmark datasets, demonstrating both the prior issues with hyperparameter selection and the effectiveness of our proposed strategy.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 58 of its 107 references that are in the map. First posted on arXiv 24 Jan 2025.
Builds on
58 of 107 references.
References · 58
- 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
- 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
- 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,282
- 2017Graph Attention Networkscited by 8,854
- 2019Fast Graph Representation Learning with PyTorch Geometriccited by 2,414
- 2019RandAugment: Practical automated data augmentation with a reduced search spacecited by 2,261
- 2020Graph Contrastive Learning with Augmentationscited by 1,203
- 2018AutoAugment: Learning Augmentation Policies from Datacited by 728
- 2019GraphSAINT: Graph Sampling Based Inductive Learning Methodcited by 618
- 2019Predicting Dynamic Embedding Trajectory in Temporal Interaction Networkscited by 443
- 2021Graph Self-Supervised Learning: A Surveycited by 289
- 2021Graph Contrastive Learning Automatedcited by 287
38 more
- 2019Fast AutoAugmentcited by 243
- 2021A Comprehensive Survey on Graph Anomaly Detection with Deep Learningcited by 230
- 2020Data Augmentation for Graph Neural Networkscited by 200
- 2021Adversarial Graph Augmentation to Improve Graph Contrastive Learningcited by 200
- 2021Self-Supervised Learning of Graph Neural Networks: A Unified Reviewcited by 184
- 2014Graph-based Anomaly Detection and Description: A Surveycited by 182
- 2020On Hyperparameter Optimization of Machine Learning Algorithms: Theory and Practicecited by 181
- 2021Anomaly Detection on Attributed Networks via Contrastive Self-Supervised Learningcited by 175
- 2019Population Based Augmentation: Efficient Learning of Augmentation Policy Schedulescited by 153
- 2020Hyperparameter Ensembles for Robustness and Uncertainty Quantificationcited by 149
- 2021Hyperparameter Optimization: Foundations, Algorithms, Best Practices and Open Challengescited by 143
- 2021Self-supervised Learning on Graphs: Contrastive, Generative,or Predictivecited by 136
- 2019Adversarial AutoAugmentcited by 111
- 2019Faster AutoAugment: Learning Augmentation Strategies using Backpropagationcited by 96
- 2017Learning to Compose Domain-Specific Transformations for Data Augmentationcited by 91
- 2020Sub-graph Contrast for Scalable Self-Supervised Graph Representation Learningcited by 90
- 2020AnomalyDAE: Dual autoencoder for anomaly detection on attributed networkscited by 81
- 2022BOND: Benchmarking Unsupervised Outlier Node Detection on Static Attributed Graphscited by 80
- 2021AutoGCL: Automated Graph Contrastive Learning via Learnable View Generatorscited by 73
- 2020Contrastive Self-supervised Learning for Graph Classificationcited by 68
- 2021Generative and Contrastive Self-Supervised Learning for Graph Anomaly Detectioncited by 67
- 2021Deep Graph-level Anomaly Detection by Glocal Knowledge Distillationcited by 67
- 2022Graph Anomaly Detection via Multi-Scale Contrastive Learning Networks with Augmented Viewcited by 54
- 2021Automated Self-Supervised Learning for Graphscited by 50
- 2019Unsupervised Model Selection for Variational Disentangled Representation Learningcited by 46
- 2020TODS: An Automated Time Series Outlier Detection Systemcited by 41
- 2020Meta-AAD: Active Anomaly Detection with Deep Reinforcement Learningcited by 34
- 2023A Survey on Graph Diffusion Models: Generative AI in Science for Molecule, Protein and Materialcited by 34
- 2016Scalable Anomaly Ranking of Attributed Neighborhoodscited by 33
- 2018LSCP: Locally Selective Combination in Parallel Outlier Ensemblescited by 31
- 2022Evidence-aware Fake News Detection with Graph Neural Networkscited by 30
- 2022Reconstruction Enhanced Multi-View Contrastive Learning for Anomaly Detection on Attributed Networkscited by 28
- 2022Hyperparameter Sensitivity in Deep Outlier Detection: Analysis and a Scalable Hyper-Ensemble Solutioncited by 24
- 2016How to Evaluate the Quality of Unsupervised Anomaly Detection Algorithms?cited by 22
- 2021Graph Contrastive Learning for Anomaly Detectioncited by 22
- 2020PyODDS: An End-to-end Outlier Detection System with Automated Machine Learningcited by 21
- 2022Automated Data Augmentations for Graph Classificationcited by 19
- 2022Label-invariant Augmentation for Semi-Supervised Graph Classificationcited by 18
50 most cited of 58 All 58 (JSON)
Built on by
Not cited yet.