Multi-scale contrastive learning sharpens anomaly detection in multivariate time series
A new contrastive learning method called MsRCL uses multi-scale patch decomposition and channel-aware attention to improve anomaly detection in multivariate ...
A new contrastive learning method called MsRCL uses multi-scale patch decomposition and channel-aware attention to improve anomaly detection in multivariate ...
Researchers in India built a transformer-based deep learning system to detect anomalies in solar-wind power plants, achieving strong reconstruction accuracy ...
Researchers have developed a hybrid deep learning architecture combining graph neural networks, Transformers, and variational autoencoders to detect and explain ...
Researchers at the University of Tabriz have developed GCAT, a graph-convolutional adversarial transformer that detects anomalies in multivariate sensor data ...
Artificial intelligence researchers have proposed a new way to make sense of one of machine learning’s fastest-growing challenges: detecting unusual ...
© 2026 Scienmag - Science Magazine
© 2026 Scienmag - Science Magazine