TY - JOUR A1 - Vu-Bac, N. A1 - Lahmer, Tom A1 - Zhuang, Xiaoying A1 - Nguyen-Thoi, T. A1 - Rabczuk, Timon T1 - A software framework for probabilistic sensitivity analysis for computationally expensive models JF - Advances in Engineering Software N2 - A software framework for probabilistic sensitivity analysis for computationally expensive models KW - Angewandte Mathematik KW - Stochastik KW - Strukturmechanik Y1 - 2016 SP - 19 EP - 31 ER - TY - JOUR A1 - Nanthakumar, S.S. A1 - Lahmer, Tom A1 - Zhuang, Xiaoying A1 - Park, Harold S. A1 - Rabczuk, Timon T1 - Topology optimization of piezoelectric nanostructures JF - Journal of the Mechanics and Physics of Solids N2 - Topology optimization of piezoelectric nanostructures KW - Angewandte Mathematik KW - Stochastik KW - Strukturmechanik Y1 - 2016 SP - 316 EP - 335 ER - TY - JOUR A1 - Nanthakumar, S.S. A1 - Lahmer, Tom A1 - Zhuang, Xiaoying A1 - Zi, Goangseup A1 - Rabczuk, Timon T1 - Detection of material interfaces using a regularized level set method in piezoelectric structures JF - Inverse Problems in Science and Engineering N2 - Detection of material interfaces using a regularized level set method in piezoelectric structures KW - Angewandte Mathematik KW - Stochastik KW - Strukturmechanik Y1 - 2016 SP - 153 EP - 176 ER - TY - JOUR A1 - Ghasemi, Hamid A1 - Brighenti, Roberto A1 - Zhuang, Xiaoying A1 - Muthu, Jacob A1 - Rabczuk, Timon T1 - Optimum fiber content and distribution in fiber-reinforced solids using a reliability and NURBS based sequential optimization approach JF - Structural and Multidisciplinary Optimization N2 - Optimum _ber content and distribution in _ber-reinforced solids using a reliability and NURBS based sequential optimization approach KW - Angewandte Mathematik KW - Strukturmechanik Y1 - 2015 SP - 99 EP - 112 ER - TY - JOUR A1 - Vu-Bac, N. A1 - Silani, Mohammad A1 - Lahmer, Tom A1 - Zhuang, Xiaoying A1 - Rabczuk, Timon T1 - A unified framework for stochastic predictions of Young's modulus of clay/epoxy nanocomposites (PCNs) JF - Computational Materials Science N2 - A unified framework for stochastic predictions of Young's modulus of clay/epoxy nanocomposites (PCNs) KW - Angewandte Mathematik KW - Stochastik KW - Strukturmechanik Y1 - 2015 SP - 520 EP - 535 ER - TY - JOUR A1 - Nanthakumar, S.S. A1 - Lahmer, Tom A1 - Zhuang, Xiaoying A1 - Zi, Goangseup A1 - Rabczuk, Timon T1 - Detection of material interfaces using a regularized level set method in piezoelectric structures JF - Inverse Problems in Science and Engineering N2 - Detection of material interfaces using a regularized level set method in piezoelectric structures KW - Angewandte Mathematik KW - Stochastik KW - Strukturmechanik Y1 - 2015 ER - TY - JOUR A1 - Vu-Bac, N. A1 - Rafiee, Roham A1 - Zhuang, Xiaoying A1 - Lahmer, Tom A1 - Rabczuk, Timon T1 - Uncertainty quantification for multiscale modeling of polymer nanocomposites with correlated parameters JF - Composites Part B: Engineering N2 - Uncertainty quantification for multiscale modeling of polymer nanocomposites with correlated parameters KW - Angewandte Mathematik KW - Stochastik KW - Strukturmechanik Y1 - 2015 SP - 446 EP - 464 ER - TY - JOUR A1 - Vu-Bac, N. A1 - Lahmer, Tom A1 - Zhang, Yancheng A1 - Zhuang, Xiaoying A1 - Rabczuk, Timon T1 - Stochastic predictions of interfacial characteristic of polymeric nanocomposites (PNCs) JF - Composites Part B Engineering N2 - Stochastic predictions of interfacial characteristic of polymeric nanocomposites (PNCs) KW - Angewandte Mathematik KW - Stochastik KW - Strukturmechanik Y1 - 2014 SP - 80 EP - 95 ER - TY - JOUR A1 - Rabczuk, Timon A1 - Guo, Hongwei A1 - Zhuang, Xiaoying A1 - Chen, Pengwan A1 - Alajlan, Naif T1 - Stochastic deep collocation method based on neural architecture search and transfer learning for heterogeneous porous media JF - Engineering with Computers N2 - We present a stochastic deep collocation method (DCM) based on neural architecture search (NAS) and transfer learning for heterogeneous porous media. We first carry out a sensitivity analysis to determine the key hyper-parameters of the network to reduce the search space and subsequently employ hyper-parameter optimization to finally obtain the parameter values. The presented NAS based DCM also saves the weights and biases of the most favorable architectures, which is then used in the fine-tuning process. We also employ transfer learning techniques to drastically reduce the computational cost. The presented DCM is then applied to the stochastic analysis of heterogeneous porous material. Therefore, a three dimensional stochastic flow model is built providing a benchmark to the simulation of groundwater flow in highly heterogeneous aquifers. The performance of the presented NAS based DCM is verified in different dimensions using the method of manufactured solutions. We show that it significantly outperforms finite difference methods in both accuracy and computational cost. KW - Maschinelles Lernen KW - Neuronales Lernen KW - Fehlerabschätzung KW - deep learning KW - neural architecture search KW - randomized spectral representation Y1 - 2022 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:gbv:wim2-20220209-45835 UR - https://link.springer.com/article/10.1007/s00366-021-01586-2 VL - 2022 SP - 1 EP - 26 PB - Springer CY - London ER - TY - JOUR A1 - Chakraborty, Ayan A1 - Anitescu, Cosmin A1 - Zhuang, Xiaoying A1 - Rabczuk, Timon T1 - Domain adaptation based transfer learning approach for solving PDEs on complex geometries JF - Engineering with Computers N2 - In machine learning, if the training data is independently and identically distributed as the test data then a trained model can make an accurate predictions for new samples of data. Conventional machine learning has a strong dependence on massive amounts of training data which are domain specific to understand their latent patterns. In contrast, Domain adaptation and Transfer learning methods are sub-fields within machine learning that are concerned with solving the inescapable problem of insufficient training data by relaxing the domain dependence hypothesis. In this contribution, this issue has been addressed and by making a novel combination of both the methods we develop a computationally efficient and practical algorithm to solve boundary value problems based on nonlinear partial differential equations. We adopt a meshfree analysis framework to integrate the prevailing geometric modelling techniques based on NURBS and present an enhanced deep collocation approach that also plays an important role in the accuracy of solutions. We start with a brief introduction on how these methods expand upon this framework. We observe an excellent agreement between these methods and have shown that how fine-tuning a pre-trained network to a specialized domain may lead to an outstanding performance compare to the existing ones. As proof of concept, we illustrate the performance of our proposed model on several benchmark problems. KW - Maschinelles Lernen KW - NURBS KW - Transfer learning KW - Domain Adaptation KW - NURBS geometry KW - Navier–Stokes equations Y1 - 2022 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:gbv:wim2-20220811-46776 UR - https://link.springer.com/article/10.1007/s00366-022-01661-2 VL - 2022 SP - 1 EP - 20 ER -