mlsanityData to 7 Oct 2026

Balanced Edge Pruning for Graph Anomaly Detection with Noisy Labels

arXiv 2407.05934 ↗First posted 8 Jul 2024cs.LG cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Graph anomaly detection (GAD) is widely applied in many areas, such as financial fraud detection and social spammer detection. Anomalous nodes in the graph not only impact their own communities but also create a ripple effect on neighbors throughout the graph structure. Detecting anomalous nodes in complex graphs has been a challenging task. While existing GAD methods assume all labels are correct, real-world scenarios often involve inaccurate annotations. These noisy labels can severely degrade GAD performance because, with anomalies representing a minority class, even a small number of mislabeled instances can disproportionately interfere with detection models. Cutting edges to mitigate the negative effects of noisy labels is a good option; however, it has both positive and negative influences and also presents an issue of weak supervision. To perform effective GAD with noisy labels, we propose REinforced Graph Anomaly Detector (REGAD) by pruning the edges of candidate nodes potentially with mistaken labels. Moreover, we design the performance feedback based on strategically crafted confident labels to guide the cutting process, ensuring optimal results. Specifically, REGAD contains two novel components. (i) A tailored policy network, which involves two-step actions to remove negative effect propagation step by step. (ii) A policy-in-the-loop mechanism to identify suitable edge removal strategies that control the propagation of noise on the graph and estimate the updated structure to obtain reliable pseudo labels iteratively. Experiments on three real-world datasets demonstrate that REGAD outperforms all baselines under different noisy ratios.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 35 of its 48 references that are in the map. First posted on arXiv 8 Jul 2024.

Builds on

35 of 48 references.

References · 35

  1. 2018Pitfalls of Graph Neural Network Evaluationcited by 994
  2. 2018Adversarial Attack on Graph Structured Datacited by 300
  3. 2019Deep Semi-Supervised Anomaly Detectioncited by 290
  4. 2020Enhancing Graph Neural Network-based Fraud Detectors against Camouflaged Fraudsterscited by 248
  5. 2021A Comprehensive Survey on Graph Anomaly Detection with Deep Learningcited by 230
  6. 2022Rethinking Graph Neural Networks for Anomaly Detectioncited by 185
  7. 2020Graph Prototypical Networks for Few-shot Learning on Attributed Networkscited by 90
  8. 2020AnomalyDAE: Dual autoencoder for anomaly detection on attributed networkscited by 81
  9. 2021NRGNN: Learning a Label Noise-Resistant Graph Neural Network on Sparsely and Noisily Labeled Graphscited by 73
  10. 2021Few-shot Network Anomaly Detection via Cross-network Meta-learningcited by 64
  11. 2023GADBench: Revisiting and Benchmarking Supervised Graph Anomaly Detectioncited by 57
  12. 2022Graph Anomaly Detection via Multi-Scale Contrastive Learning Networks with Augmented Viewcited by 54
23 more
  1. 2023Truncated Affinity Maximization: One-class Homophily Modeling for Graph Anomaly Detectioncited by 43
  2. 2023Towards Self-Interpretable Graph-Level Anomaly Detectioncited by 41
  3. 2022Reinforced Causal Explainer for Graph Neural Networkscited by 34
  4. 2022Cross-Domain Graph Anomaly Detection via Anomaly-aware Contrastive Alignmentcited by 31
  5. 2024Alleviating Structural Distribution Shift in Graph Anomaly Detectioncited by 31
  6. 2019Weak Supervision for Fake News Detection via Reinforcement Learningcited by 30
  7. 2022Robust Training of Graph Neural Networks via Noise Governancecited by 29
  8. 2020Causal Understanding of Fake News Dissemination on Social Mediacited by 27
  9. 2022Challenges and Opportunities in Deep Reinforcement Learning with Graph Neural Networks: A Comprehensive review of Algorithms and Applicationscited by 27
  10. 2023Learning Strong Graph Neural Networks with Weak Informationcited by 25
  11. 2023Weakly Supervised Anomaly Detection: A Surveycited by 22
  12. 2024Entity Alignment with Noisy Annotations from Large Language Modelscited by 22
  13. 2021Noise-robust Graph Learning by Estimating and Leveraging Pairwise Interactionscited by 20
  14. 2023GraphSR: A Data Augmentation Algorithm for Imbalanced Node Classificationcited by 19
  15. 2022Reinforcement learning on graphs: A surveycited by 18
  16. 2022ADMoE: Anomaly Detection with Mixture-of-Experts from Noisy Labelscited by 15
  17. 2024Modality-Aware Integration with Large Language Models for Knowledge-based Visual Question Answeringcited by 15
  18. 2021Relevance-Aware Anomalous Users Detection in Social Network via Graph Neural Networkcited by 12
  19. 2023Improving Generalizability of Graph Anomaly Detection Models via Data Augmentationcited by 11
  20. 2020Learning to Deceive Knowledge Graph Augmented Models via Targeted Perturbationcited by 10
  21. 2021A Generic Graph Sparsification Framework using Deep Reinforcement Learningcited by 8
  22. 2021Robust Graph Meta-learning for Weakly-supervised Few-shot Node Classificationcited by 2
  23. 2022Label Noise-Resistant Mean Teaching for Weakly Supervised Fake News Detectioncited by 1

Built on by

Not cited yet.