Prior laundering: learned priors with inherited, undetectable overconfidence

The document is an arXiv statistical machine learning feed containing newly announced research papers on learned priors, simulation-based empirical Bayes, LLM watermark detection, optimization convergence, tensor decomposition, predictive maintenance, time-series forecasting, causal inference, dynamical systems, correlation-matrix generation, and related topics. The content is academic and does not describe an exploitable vulnerability, security incident, malware, or threat actor activity. The first paper discusses inherited overconfidence and uncertainty miscalibration in learned generative “