Tag: Anomaly Detection

A new framework called AMG improves unsupervised anomaly detection in complex attributed multiplex heterogeneous networks by integrating multi-channel graph convolutional ...

A federated unlearning framework eliminates target drone data from intrusion detection models while cutting communication rounds by 95 percent on ...

A new editorial collection documents how artificial intelligence is reshaping particle accelerator science, from surrogate models already in practical use ...

A new study proposes a closed-loop AI framework that couples adaptive anomaly detection, probabilistic risk scoring, machine-learning measurement selection, game-theoretic ...

An AI framework combining supervised learning, autoencoder anomaly detection and a multi-agent system detected known and unseen faults in a ...

A new framework discretizes time-series data and splits it between global rule-based and Transformer-based local detectors, reporting higher F1 scores ...

Researchers have proposed PSI-SAD-DA, a post-selection inference framework that provides valid p-values and false positive rate control for semi-supervised anomaly ...

Most intrusion detection research is tested on artificially balanced datasets, where attack and benign records appear in equal numbers. A ...

A new special issue of Data Science and Engineering collects eleven peer-reviewed papers on data spaces, data governance, and data-driven ...

Researchers introduced EVENET, a foundation model pretrained on 500 million simulated collision events that outperforms task-specific baselines across new-physics searches, ...

A new graph neural network framework called EGNN expands rare attack samples and uses threshold-controlled pseudo-labeling to improve intrusion detection ...

Internal implants and anomaly detection algorithms flagged significant activity changes in rabies-infected foxes up to two weeks before death, while ...

A new study reports that Random Forest and Gaussian and Sigmoid kernel Support Vector Machines, combined with systematic preprocessing, achieved ...

The European Spallation Source is developing a strategy and pilot projects to apply artificial intelligence and machine learning across its ...

A new EPJ Data Science study models 27 million AIS signals from the 2018 Eastern Black Sea to link prolonged ...

Two unsupervised machine-learning methods trained on SuperKEKB beam position monitor data matched each other's detection performance and flagged nine unreported ...

A new open-access study presents EUC_SOH, a monitoring system that derives battery and power-electronics health indicators for electric unicycles from ...

A new simulation study presents GuardBox, a framework combining an LSTM autoencoder and random forest classifier to detect and attribute ...

A new contrastive learning method called MsRCL uses multi-scale patch decomposition and channel-aware attention to improve anomaly detection in multivariate ...

A conditional diffusion model reached an F1-score of 0.975 when detecting abnormal data in a 90-day, 18-channel bridge monitoring dataset ...