Model-Free Local Recalibration of Neural Networks

arXiv:2403.05756v1 Announce Type: cross
Abstract: Artificial neural networks (ANNs) are highly flexible predictive models. However, reliably quantifying uncertainty for their predictions is a continuing challenge. There has been much recent work on "recalibration" of predictive distributions for ANNs, so that forecast probabilities for events of interest are consistent with certain frequency evaluations of them. Uncalibrated probabilistic forecasts are of limited use for many important decision-making tasks. To address this issue, we propose a localized recalibration of ANN predictive distributions using …

abstract anns artificial artificial neural networks arxiv challenge consistent cs.lg events forecast free however networks neural networks predictions predictive predictive models stat.me them type uncertainty work