SPECTRA: State-Space Exogenous Context and Temporal-Frequency Resolution Architecture for Probabilistic Energy Forecasting

Collection of recent machine-learning and statistics research papers (arXiv) covering probabilistic energy forecasting (SPECTRA), automatic knot selection for spline/GAMs, transformer-based diffusion models for hydrological time-series imputation and forecasting, Generative Bayesian Filtering (CVAE + score-based measurement updates) for state estimation, ConfidenceBench (benchmarking verbalized confidence calibration in LLMs), CLOE (Christoffel Loss Autoencoder) for anomaly detection, theoretical work on Fisher widths, weight-tied/looped transformer algorithmic behavior, and Gaussian-processs/