Learning Track

Each chapter pairs a theory document with a companion notebook (notebooks/).

Chapter	Topic
ch01	Defining the TSAD problem: anomaly types (point/contextual/collective), contamination
ch02	Gen1 statistics: control charts, T²/SPE derivation, STL
ch03	Gen2 classical ML: IForest, Matrix Profile / discords
ch04	Gen3 deep learning: reconstruction vs forecasting, VAE, over-generalization
ch05	Gen4: graphs (GDN) and attention (association discrepancy)
ch06	Gen5: SSM/Mamba, foundation models and zero-shot
ch07	Evaluation methodology (most important): PA inflation, VUS/affiliation
ch08	Thresholding and decisions: EVT (SPOT/DSPOT), conformal
ch09	Industrial practice: a semiconductor FDC/ATE perspective, BYOD
ch10	How to read the benchmark + full bibliography

Suggested paths: practitioners start with ch01 → ch07 → ch08 → ch09; researchers read in order.