Credal Bayesian Deep Learning

arXiv:2302.09656v4 Announce Type: replace
Abstract: Uncertainty quantification and robustness to distribution shifts are important goals in machine learning and artificial intelligence. Although Bayesian Neural Networks (BNNs) allow for uncertainty in the predictions to be assessed, different sources of uncertainty are indistinguishable. We present Credal Bayesian Deep Learning (CBDL). Heuristically, CBDL allows to train an (uncountably) infinite ensemble of BNNs, using only finitely many elements. This is possible thanks to prior and likelihood finitely generated credal sets (FGCSs), a concept from …

abstract artificial artificial intelligence arxiv bayesian bayesian deep learning credal cs.lg deep learning distribution intelligence machine machine learning networks neural networks predictions quantification robustness stat.ml train type uncertainty