Anomaly Detection with Variance Stabilized Density Estimation

Amit Rozner, Barak Battash, Henry Li, Lior Wolf, Ofir Lindenbaum
Proceedings of the Fortieth Conference on Uncertainty in Artificial Intelligence, PMLR 244:3121-3137, 2024.

Abstract

We propose a modified density estimation problem that is highly effective for detecting anomalies in tabular data. Our approach assumes that the density function is relatively stable (with lower variance) around normal samples. We have verified this hypothesis empirically using a wide range of real-world data. Then, we present a variance-stabilized density estimation problem for maximizing the likelihood of the observed samples while minimizing the variance of the density around normal samples. To obtain a reliable anomaly detector, we introduce a spectral ensemble of autoregressive models for learning the variance-stabilized distribution. We have conducted an extensive benchmark with 52 datasets, demonstrating that our method leads to state-of-the-art results while alleviating the need for data-specific hyperparameter tuning. Finally, we have used an ablation study to demonstrate the importance of each of the proposed components, followed by a stability analysis evaluating the robustness of our model.

Cite this Paper

BibTeX
@InProceedings{pmlr-v244-rozner24a, title = {Anomaly Detection with Variance Stabilized Density Estimation}, author = {Rozner, Amit and Battash, Barak and Li, Henry and Wolf, Lior and Lindenbaum, Ofir}, booktitle = {Proceedings of the Fortieth Conference on Uncertainty in Artificial Intelligence}, pages = {3121--3137}, year = {2024}, editor = {Kiyavash, Negar and Mooij, Joris M.}, volume = {244}, series = {Proceedings of Machine Learning Research}, month = {15--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v244/main/assets/rozner24a/rozner24a.pdf}, url = {https://proceedings.mlr.press/v244/rozner24a.html}, abstract = {We propose a modified density estimation problem that is highly effective for detecting anomalies in tabular data. Our approach assumes that the density function is relatively stable (with lower variance) around normal samples. We have verified this hypothesis empirically using a wide range of real-world data. Then, we present a variance-stabilized density estimation problem for maximizing the likelihood of the observed samples while minimizing the variance of the density around normal samples. To obtain a reliable anomaly detector, we introduce a spectral ensemble of autoregressive models for learning the variance-stabilized distribution. We have conducted an extensive benchmark with 52 datasets, demonstrating that our method leads to state-of-the-art results while alleviating the need for data-specific hyperparameter tuning. Finally, we have used an ablation study to demonstrate the importance of each of the proposed components, followed by a stability analysis evaluating the robustness of our model.} }
Endnote
%0 Conference Paper %T Anomaly Detection with Variance Stabilized Density Estimation %A Amit Rozner %A Barak Battash %A Henry Li %A Lior Wolf %A Ofir Lindenbaum %B Proceedings of the Fortieth Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2024 %E Negar Kiyavash %E Joris M. Mooij %F pmlr-v244-rozner24a %I PMLR %P 3121--3137 %U https://proceedings.mlr.press/v244/rozner24a.html %V 244 %X We propose a modified density estimation problem that is highly effective for detecting anomalies in tabular data. Our approach assumes that the density function is relatively stable (with lower variance) around normal samples. We have verified this hypothesis empirically using a wide range of real-world data. Then, we present a variance-stabilized density estimation problem for maximizing the likelihood of the observed samples while minimizing the variance of the density around normal samples. To obtain a reliable anomaly detector, we introduce a spectral ensemble of autoregressive models for learning the variance-stabilized distribution. We have conducted an extensive benchmark with 52 datasets, demonstrating that our method leads to state-of-the-art results while alleviating the need for data-specific hyperparameter tuning. Finally, we have used an ablation study to demonstrate the importance of each of the proposed components, followed by a stability analysis evaluating the robustness of our model.
APA
Rozner, A., Battash, B., Li, H., Wolf, L. & Lindenbaum, O.. (2024). Anomaly Detection with Variance Stabilized Density Estimation. Proceedings of the Fortieth Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 244:3121-3137 Available from https://proceedings.mlr.press/v244/rozner24a.html.

Related Material