Example of Anomaly Detection using Convolutional Variational Auto-Encoder (CVAE)
Example of Anomaly Detection using Convolutional Variational Auto-Encoder (CVAE) [TensorFlow 1.x] [TensorFlow 2.x].
Simplified VAE architecture.
'Class-1' is defined as normal and the others are defined as abnormal.
MNIST | Fashion-MNIST | |
---|---|---|
Reconstruciton of training | ||
Latent of training | ||
Latent walk | ||
Latent of test | ||
Histogram of test | ||
AUROC | 0.997 | 0.980 |
[1] Kingma, D. P., & Welling, M. (2013). Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114.
[2] Kullback Leibler divergence. Wikipedia