Wiley-Blackwell publishers announced that the 2019 Hojjat Adeli Award for Innovation in Computing was awarded to Xihaier Luo and Ahsan Kareem for their paper titled Deep convolutional neural networks for uncertainty propagation in random fields. The award is granted annually to the most innovative paper published in the previous volume.
Xihaier Luo is a graduate student at University of Notre Dame involved in the NatHaz Modeling Laboratory, and Ahsan Kareem is the co-PI lead at the NSF NHERI SimCenter for wind hazards in addition to being the Robert M. Moran Professor of Engineering in the Department of Civil & Environmental Engineering & Earth Sciences at the University of Notre Dame.
Kareem comments that this paper utilizing convolutional neural networks offers a reliable and robust surrogate model for systems often constrained by the dimensionality of the problem. To assess the model performance, uncertainty quantification is carried out in a continuum mechanics benchmark problem. The results suggest that the proposed model is capable of directly inferring a wide variety of I/O mapping relationships, and the model also characterizes the statistical properties of the output fields comparable to Monte Carlo estimates. This opens avenues for mapping complex relationships that are currently mathematically intractable, like relationships between turbulence to buffeting effects and beyond.
Access the paper from the Wiley Online Library.
Citation: Xihaier Luo and Ahsan Kareem (2019), Deep convolutional neural networks for uncertainty propagation in random fields, Computer-Aided Civil and Infrastructure Engineering, 34:12, 1041-1054.
