Research map / Image generation / GANs
GAN Ensemble for Anomaly Detection
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Replace single f-AnoGAN with a 3-generator, 3-discriminator ensemble | f-AnoGAN · MNIST (average) | AUROC: 0.799 -> 0.857 | Helped |
| Replace single GANomaly with a 3-generator, 3-discriminator ensemble | GANomaly · MNIST (average) | AUROC: 0.673 -> 0.830 | Helped |
| Replace single f-AnoGAN with a 3-generator, 3-discriminator ensemble | f-AnoGAN · CIFAR-10 | AUROC | Helped |
| Replace single GANomaly with a 3-generator, 3-discriminator ensemble | GANomaly · CIFAR-10 | AUROC | Helped |
| Replace single Skip-GANomaly with a 3-generator, 3-discriminator ensemble | Skip-GANomaly · CIFAR-10 | AUROC: 0.731 -> 0.868 | Helped |
| Replace single Skip-GANomaly with a 3-generator, 3-discriminator ensemble | Skip-GANomaly · OCT (overall) | AUROC: 0.766 -> 0.869 | Helped |
| Replace single f-AnoGAN with a 3-generator, 3-discriminator ensemble | f-AnoGAN · KDD99 | F1: 0.960 -> 0.979 | Helped |
| Replace single EGBAD with a 3-generator, 3-discriminator ensemble | EGBAD · KDD99 | F1: 0.939 -> 0.966 | Helped |
| Increase ensemble size from 1 to 3 (ablation) | GAN ensembles · CIFAR-10 (ship anomalous class) | AUROC: 35.9 % | Helped |
| Increase ensemble size from 3 to 7 (ablation) | GAN ensembles · CIFAR-10 (ship anomalous class) | AUROC | No clear effect |
| Increase the relative weight of discriminative loss (ablation) | GAN ensemble anomaly score · CIFAR-10 | detection performance | Helped |
Closest papers by meaning
- Old Is Gold: Redefining the Adversarially Learned One-Class Classifier Training Paradigm (CVPR 2020 · similarity 0.90)
- Self-Diagnosing GAN: Diagnosing Underrepresented Samples in Generative Adversarial Networks (NeurIPS 2021 · similarity 0.87)
- Adversarially Learned One-Class Classifier for Novelty Detection (CVPR 2018 · similarity 0.87)
- Adversarially Robust Anomaly Detection through Spurious Negative Pair Mitigation (ICLR 2025 · similarity 0.86)
- Contamination-Resilient Anomaly Detection via Adversarial Learning on Partially-Observed Normal and Anomalous Data (ICML 2024 · similarity 0.86)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.