Table 1 Effect of the sample size \(N_{s}\) on the maximum relative error in the \(L^{2}\)-norm for the mean and standard deviation obtained from the NIROM-CAEs (\(L_{x}=50\), \(L_{t}=10\)) and POD-ANN (\(\epsilon _{s}=\epsilon _{t}=10^{-8}\)) approaches. Errors are computed with respect to the LHS reference solution (with \(N_{s}=2000\) realizations)

From: Reduced-order modeling for stochastic large-scale and time-dependent flow problems using deep spatial and temporal convolutional autoencoders