Can AI produce valid risk forecasts?
A pretrained model always returns a number that looks like a Value at Risk; this lecture is about telling whether it answered the question you asked.
Classical baselines, a time-series foundation model and language models are put through the same backtest across several markets — the answer splits by model, and that is the finding.
Daniel Traian Pele is a Prof. dr. Department of Statistics and Econometrics Faculty of Cybernetics, Statistics and Economic Informatics, The Bucharest University of Economic Studies. https://scholar.google.com/citations?user=tN32HYcAAAAJ&hl=en