Efficient GAN-Based Anomaly Detection
About
Generative adversarial networks (GANs) are able to model the complex highdimensional distributions of real-world data, which suggests they could be effective for anomaly detection. However, few works have explored the use of GANs for the anomaly detection task. We leverage recently developed GAN models for anomaly detection, and achieve state-of-the-art performance on image and network intrusion datasets, while being several hundred-fold faster at test time than the only published GAN-based method.
Houssam Zenati, Chuan Sheng Foo, Bruno Lecouat, Gaurav Manek, Vijay Ramaseshan Chandrasekhar• 2018
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Anomaly Detection | MVTecAD (test) | Bottle Score68 | 55 | |
| Anomaly Detection | MVTec AD | AUROC0.6 | 35 | |
| Anomaly Detection | MVTec Anomaly Detection 1.0 (test) | PRO (Carpet)0.52 | 27 | |
| Anomaly Detection | MNIST | AUPRC (Digit 1)0.287 | 7 | |
| Anomaly Detection | MNIST Heldout Digit 1 (test) | AUPRC28.7 | 7 | |
| Anomaly Detection | MNIST Heldout Digit 4 1 (test) | AUPRC0.443 | 7 | |
| Anomaly Detection | MNIST Heldout Digit 5 1 (test) | AUPRC0.514 | 7 | |
| Anomaly Detection | MNIST Heldout Digit 7 1 (test) | AUPRC0.347 | 7 | |
| Anomaly Detection | MNIST Heldout Digit 9 1 (test) | AUPRC30.7 | 7 |
Showing 9 of 9 rows