TY - JOUR
T1 - Anomaly Detection with Variance Stabilized Density Estimation
AU - Rozner, Amit
AU - Battash, Barak
AU - Li, Henry
AU - Wolf, Lior
AU - Lindenbaum, Ofir
N1 - Publisher Copyright:
© 2024 Proceedings of Machine Learning Research. All rights reserved.
PY - 2024
Y1 - 2024
N2 - 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.
AB - 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.
UR - http://www.scopus.com/inward/record.url?scp=85212216200&partnerID=8YFLogxK
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.conferencearticle???
AN - SCOPUS:85212216200
SN - 2640-3498
VL - 244
SP - 3121
EP - 3137
JO - Proceedings of Machine Learning Research
JF - Proceedings of Machine Learning Research
T2 - 40th Conference on Uncertainty in Artificial Intelligence, UAI 2024
Y2 - 15 July 2024 through 19 July 2024
ER -