Reliable uncertainty with cheaper neural network ensembles: a case study in industrial parts classification

arXiv:2403.10182v1 Announce Type: new
Abstract: In operations research (OR), predictive models often encounter out-of-distribution (OOD) scenarios where the data distribution differs from the training data distribution. In recent years, neural networks (NNs) are gaining traction in OR for their exceptional performance in fields such as image classification. However, NNs tend to make confident yet incorrect predictions when confronted with OOD data. Uncertainty estimation offers a solution to overconfident models, communicating when the output should (not) be trusted. Hence, reliable uncertainty …

abstract arxiv case case study classification cs.lg data distribution fields image industrial network networks neural network neural networks nns operations performance predictive predictive models research stat.ml study training training data type uncertainty