Shapley-value anomaly tests equal simpler single-term tests for independent sensors but differ for correlated bivariate Gaussians, with strict superiority or inferiority depending on correlation sign.
Deep learning for anomaly detection
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
verdicts
UNVERDICTED 9representative citing papers
A rule-based strikingness measure is added to TKGR metrics to weight rare events higher, revealing that models weaken on striking events and ensemble gains come mostly from trivial fits.
uLEAD-TabPFN detects anomalies in tabular data by scoring violations of conditional dependencies estimated via frozen PFNs with uncertainty awareness, achieving top average rank and up to 20% ROC-AUC gains on high-dimensional ADBench datasets.
Diffusion models via DDPM work for anomaly detection but are slow; the proposed DTE method estimates diffusion time distribution analytically and with a neural net to deliver faster inference while outperforming DDPM on ADBench for unsupervised and semi-supervised settings.
A single fixed term in the Shapley value yields the same anomaly localization error probability as the full calculation for independent sensor observations, supported by a proof.
Tensor Train compression algorithms detect anomalies by maintaining normal data structure and deleting anomalous structure, tested on digits, faces, and cyber-attack datasets.
Proposes the TRIAD framework that treats multi-turn multimodal attacks as continuous trajectories and uses structural anomaly detection, regularized Mahalanobis distance, topological acceleration, and a time-varying Cox model with Bayesian HMM feedback to predict and bound expected time-to-failure.
ESN-DAGMM adapts DAGMM with an ESN layer for temporal modeling and reports 269.59% better average clustering quality than baselines on 10% of an O-RAN video-streaming dataset.
Genetic algorithm feature selection with Extra Trees reduces PMU features from 112 to an average of 27.4 while raising macro-F1 from 0.9118 to 0.9212 and ROC-AUC from 0.9791 to 0.9837 on the MSU/ORNL dataset.
citing papers explorer
-
uLEAD-TabPFN: Uncertainty-aware Dependency-based Anomaly Detection with TabPFN
uLEAD-TabPFN detects anomalies in tabular data by scoring violations of conditional dependencies estimated via frozen PFNs with uncertainty awareness, achieving top average rank and up to 20% ROC-AUC gains on high-dimensional ADBench datasets.
-
On Diffusion Modeling for Anomaly Detection
Diffusion models via DDPM work for anomaly detection but are slow; the proposed DTE method estimates diffusion time distribution analytically and with a neural net to deliver faster inference while outperforming DDPM on ADBench for unsupervised and semi-supervised settings.
-
On Using the Shapley Value for Anomaly Localization: A Statistical Investigation
A single fixed term in the Shapley value yields the same anomaly localization error probability as the full calculation for independent sensor observations, supported by a proof.
-
Anomaly Detection from a Tensor Train Perspective
Tensor Train compression algorithms detect anomalies by maintaining normal data structure and deleting anomalous structure, tested on digits, faces, and cyber-attack datasets.
-
Cyber-Physical Anomaly Detection in IoT-Enabled Smart Grids Using Machine Learning and Metaheuristic Feature Optimization
Genetic algorithm feature selection with Extra Trees reduces PMU features from 112 to an average of 27.4 while raising macro-F1 from 0.9118 to 0.9212 and ROC-AUC from 0.9791 to 0.9837 on the MSU/ORNL dataset.