[New Paper] Ensemble-based Uncertainty Quantification and Decomposition of Probabilistic Surrogate Models Using Bayesian Neural Networks
Published:
Our paper “Ensemble-based Uncertainty Quantification and Decomposition of Probabilistic Surrogate Models Using Bayesian Neural Networks” has been published in Structural Safety (available online 13 March 2026).
The study presents a new framework for decomposing predictive uncertainty in ensembles of Bayesian neural networks, systematically separating irreducible randomness from reducible model-form and data-induced uncertainties.
Citation: Jeon, J., Song, J., & Kwon, O. (2026). “Ensemble-based Uncertainty Quantification and Decomposition of Probabilistic Surrogate Models Using Bayesian Neural Networks.” Structural Safety. https://doi.org/10.1016/j.strusafe.2026.102698
