Identification and Inference in Nonlinear Dynamic Network Models

Collection of recent stat-ml arXiv preprints covering methods for identification and inference in nonlinear dynamic networks; semi-parametric state-space models for learning nonlinear regime transitions; a structured source-wise adaptive diffusion framework for blind source separation; a new time-sensitive detection metric (Hiremath Early Detection, HED) with applications to cyber-physical security and anomaly detection; anticipatory generative jump-diffusions for discontinuous path synthesis; individual-heterogeneous sub-Gaussian mixture models and spectral clustering; machine‑learning‑asss­-