Unsupervised anomaly detection with adversarial mirrored autoencoders

Gowthami Somepalli, Yexin Wu, Yogesh Balaji, Bhanukiran Vinzamuri, Soheil Feizi
Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence, PMLR 161:365-375, 2021.

Abstract

Detecting out-of-distribution (OOD) samples is of paramount importance in all Machine Learning applications. Deep generative modeling has emerged as a dominant paradigm to model complex data distributions without labels. However, prior work has shown that generative models tend to assign higher likelihoods to OOD samples compared to the data distribution on which they were trained. First, we propose Adversarial Mirrored Autoencoder (AMA), a variant of Adversarial Autoencoder, which uses a mirrored Wasserstein loss in the discriminator to enforce better semantic-level reconstruction. We also propose a latent space regularization to learn a compact manifold for in-distribution samples. The use of AMA produces better feature representations that improve anomaly detection performance. Second, we put forward an alternative measure of anomaly score to replace the reconstruction-based metric which has been traditionally used in generative model-based anomaly detection methods. Our method outperforms the current state-of-the-art methods for anomaly detection on several OOD detection benchmarks.

Cite this Paper


BibTeX
@InProceedings{pmlr-v161-somepalli21a, title = {Unsupervised anomaly detection with adversarial mirrored autoencoders}, author = {Somepalli, Gowthami and Wu, Yexin and Balaji, Yogesh and Vinzamuri, Bhanukiran and Feizi, Soheil}, booktitle = {Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence}, pages = {365--375}, year = {2021}, editor = {de Campos, Cassio and Maathuis, Marloes H.}, volume = {161}, series = {Proceedings of Machine Learning Research}, month = {27--30 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v161/somepalli21a/somepalli21a.pdf}, url = {https://proceedings.mlr.press/v161/somepalli21a.html}, abstract = {Detecting out-of-distribution (OOD) samples is of paramount importance in all Machine Learning applications. Deep generative modeling has emerged as a dominant paradigm to model complex data distributions without labels. However, prior work has shown that generative models tend to assign higher likelihoods to OOD samples compared to the data distribution on which they were trained. First, we propose Adversarial Mirrored Autoencoder (AMA), a variant of Adversarial Autoencoder, which uses a mirrored Wasserstein loss in the discriminator to enforce better semantic-level reconstruction. We also propose a latent space regularization to learn a compact manifold for in-distribution samples. The use of AMA produces better feature representations that improve anomaly detection performance. Second, we put forward an alternative measure of anomaly score to replace the reconstruction-based metric which has been traditionally used in generative model-based anomaly detection methods. Our method outperforms the current state-of-the-art methods for anomaly detection on several OOD detection benchmarks.} }
Endnote
%0 Conference Paper %T Unsupervised anomaly detection with adversarial mirrored autoencoders %A Gowthami Somepalli %A Yexin Wu %A Yogesh Balaji %A Bhanukiran Vinzamuri %A Soheil Feizi %B Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2021 %E Cassio de Campos %E Marloes H. Maathuis %F pmlr-v161-somepalli21a %I PMLR %P 365--375 %U https://proceedings.mlr.press/v161/somepalli21a.html %V 161 %X Detecting out-of-distribution (OOD) samples is of paramount importance in all Machine Learning applications. Deep generative modeling has emerged as a dominant paradigm to model complex data distributions without labels. However, prior work has shown that generative models tend to assign higher likelihoods to OOD samples compared to the data distribution on which they were trained. First, we propose Adversarial Mirrored Autoencoder (AMA), a variant of Adversarial Autoencoder, which uses a mirrored Wasserstein loss in the discriminator to enforce better semantic-level reconstruction. We also propose a latent space regularization to learn a compact manifold for in-distribution samples. The use of AMA produces better feature representations that improve anomaly detection performance. Second, we put forward an alternative measure of anomaly score to replace the reconstruction-based metric which has been traditionally used in generative model-based anomaly detection methods. Our method outperforms the current state-of-the-art methods for anomaly detection on several OOD detection benchmarks.
APA
Somepalli, G., Wu, Y., Balaji, Y., Vinzamuri, B. & Feizi, S.. (2021). Unsupervised anomaly detection with adversarial mirrored autoencoders. Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 161:365-375 Available from https://proceedings.mlr.press/v161/somepalli21a.html.

Related Material