id	author	title	date	pages	extension	mime	words	sentence	flesch	summary	cache	txt
cana-6095	Shobhankumar D.M	Uncertainty Quantification in Numerical Methods Using Deep Bayesian Neural Networks	2023	11	.pdf	application/pdf	5250	380	13	By treating network weights as probability distributions rather than point estimates, BNNs naturally capture both aleatoric uncertainty (inherent randomness in data) and epistemic uncertainty (model uncertainty due to limited data) 1,8. For regression tasks, this can be achieved by modeling heteroscedastic noise https://www.sciencedirect.com/science/article/abs/pii/S016794731930163X https://www.sciencedirect.com/science/article/abs/pii/S0045782521004102 https://www.sciencedirect.com/science/article/abs/pii/S0021999122009652 https://arxiv.org/abs/2102.06559 https://www.sciencedirect.com/science/article/abs/pii/S0045782521004102 https://arxiv.org/abs/2309.16314 https://www.sciencedirect.com/science/article/abs/pii/S016794731930163X https://www.sciencedirect.com/science/article/abs/pii/S0045782521004102 https://www.sciencedirect.com/science/article/abs/pii/S016794731930163X https://arxiv.org/abs/2412.08776 https://www.sciencedirect.com/science/article/abs/pii/S016794731930163X https://www.sciencedirect.com/science/article/abs/pii/S016794731930163X Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 30 No. 3 (2023) 60 https://internationalpubls.com (aleatoric) while capturing parameter uncertainty (epistemic).	cache/cana-6095.pdf	txt/cana-6095.txt