Anomaly Detection in IEC-61850 GOOSE Networks: Evaluating Unsupervised and Temporal Learning for Real-Time Intrusion Detection

This collection of papers covers applied security research across industrial control systems (ICS), cloud sovereignty, privacy-preserving biometrics, mobile backend security, adversarial and poisoning threats to ML-based detectors, agentic reverse-engineering limits, mutual TLS usability, and safe agent communication languages. Key operational findings: (1) For IEC-61850 GOOSE networks, unsupervised temporal models meet sub-4ms real-time constraints while retaining good detection (best GRU: F1=0.8737 at 1.118ms); supervised Random Forest performed best (F1=0.9516) but is too slow (21.8ms). (2)