English

Anomaly Detection with Variance Stabilized Density Estimation

Machine Learning 2024-05-09 v2 Artificial Intelligence

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.

Keywords

Cite

@article{arxiv.2306.00582,
  title  = {Anomaly Detection with Variance Stabilized Density Estimation},
  author = {Amit Rozner and Barak Battash and Henry Li and Lior Wolf and Ofir Lindenbaum},
  journal= {arXiv preprint arXiv:2306.00582},
  year   = {2024}
}

Comments

9 pages, 7 figures

R2 v1 2026-06-28T10:53:12.505Z