no code implementations • 23 Feb 2024 • Himanshu Sharma, Lukáš Novák, Michael D. Shields
We present a novel physics-constrained polynomial chaos expansion as a surrogate modeling method capable of performing both scientific machine learning (SciML) and uncertainty quantification (UQ) tasks.
no code implementations • 30 Jan 2024 • Dimitris G. Giovanis, Dimitrios Loukrezis, Ioannis G. Kevrekidis, Michael D. Shields
To this end, we employ Principal Geodesic Analysis on the Grassmann manifold of the response to identify a set of disjoint principal geodesic submanifolds, of possibly different dimension, that captures the variation in the data.
no code implementations • 10 Jan 2024 • Denny Thaler, Somayajulu L. N. Dhulipala, Franz Bamer, Bernd Markert, Michael D. Shields
We present a new Subset Simulation approach using Hamiltonian neural network-based Monte Carlo sampling for reliability analysis.
no code implementations • 4 Sep 2023 • Lukáš Novák, Himanshu Sharma, Michael D. Shields
This paper presents a novel methodology for the construction of physics-informed polynomial chaos expansions (PCE) that combines the conventional experimental design with additional constraints from the physics of the model.
no code implementations • 27 Aug 2023 • Mohit Chauhan, Mariel Ojeda-Tuz, Ryan Catarelli, Kurtis Gurley, Dimitrios Tsapetis, Michael D. Shields
We propose a novel strategy for active learning that focuses on resolving the main effects of the Gaussian process (associated with the numerator of the Sobol index) and compare this with existing strategies based on convergence in the total variance (the denominator of the Sobol index).
no code implementations • 29 Jun 2023 • Himanshu Sharma, Jim A. Gaffney, Dimitrios Tsapetis, Michael D. Shields
Since there are inherent uncertainties in the calibration data (parametric uncertainty) and the assumed functional EOS form (model uncertainty), it is essential to perform uncertainty quantification (UQ) to improve confidence in the EOS predictions.
1 code implementation • 15 Apr 2023 • Katiana Kontolati, Somdatta Goswami, George Em Karniadakis, Michael D. Shields
Operator regression provides a powerful means of constructing discretization-invariant emulators for partial-differential equations (PDEs) describing physical systems.
no code implementations • 31 Jan 2023 • Lukáš Novák, Michael D. Shields, Václav Sadílek, Miroslav Vořechovský
The numerical results show the superiority of the DAL-PCE in comparison to (i) a single global polynomial chaos expansion and (ii) the recently proposed stochastic spectral embedding (SSE) method developed as an accurate surrogate model and which is based on a similar domain decomposition process.
no code implementations • 7 Dec 2022 • Promit Chakroborty, Somayajulu L. N. Dhulipala, Yifeng Che, Wen Jiang, Benjamin W. Spencer, Jason D. Hales, Michael D. Shields
The multi-fidelity surrogate is assembled by first applying a Gaussian process correction to each low-fidelity model and assigning a model probability based on the model's local predictive accuracy and cost.
1 code implementation • 19 Sep 2022 • Somayajulu L. N. Dhulipala, Yifeng Che, Michael D. Shields
We propose the use of HNNs for performing Bayesian inference efficiently without requiring numerous posterior gradients.
1 code implementation • 12 Aug 2022 • Somayajulu L. N. Dhulipala, Yifeng Che, Michael D. Shields
Compared to traditional NUTS, L-HNNs in NUTS with online error monitoring required 1--2 orders of magnitude fewer numerical gradients of the target density and improved the effective sample size (ESS) per gradient by an order of magnitude.
1 code implementation • 20 Apr 2022 • Somdatta Goswami, Katiana Kontolati, Michael D. Shields, George Em Karniadakis
Transfer learning (TL) enables the transfer of knowledge gained in learning to perform one task (source) to a related but different task (target), hence addressing the expense of data acquisition and labeling, potential computational power limitations, and dataset distribution mismatches.
1 code implementation • 9 Mar 2022 • Katiana Kontolati, Somdatta Goswami, Michael D. Shields, George Em Karniadakis
In contrast, an even highly over-parameterized DeepONet leads to better generalization for both smooth and non-smooth dynamics.
no code implementations • 9 Feb 2022 • Katiana Kontolati, Dimitrios Loukrezis, Dimitris G. Giovanis, Lohit Vandanapu, Michael D. Shields
Constructing surrogate models for uncertainty quantification (UQ) on complex partial differential equations (PDEs) having inherently high-dimensional $\mathcal{O}(10^{\ge 2})$ stochastic inputs (e. g., forcing terms, boundary conditions, initial conditions) poses tremendous challenges.
no code implementations • 6 Jan 2022 • Somayajulu L. N. Dhulipala, Michael D. Shields, Promit Chakroborty, Wen Jiang, Benjamin W. Spencer, Jason D. Hales, Vincent M. Laboure, Zachary M. Prince, Chandrakanth Bolisetti, Yifeng Che
However, TRISO failure probabilities are small and the associated computational models are expensive.
no code implementations • 29 Oct 2021 • Kshitiz Upadhyay, Dimitris G. Giovanis, Ahmed Alshareef, Andrew K. Knutsen, Curtis L. Johnson, Aaron Carass, Philip V. Bayly, Michael D. Shields, K. T. Ramesh
This framework is demonstrated on a 2D subject-specific head model, where the goal is to quantify uncertainty in the simulated strain fields (i. e., output), given variability in the material properties of different brain substructures (i. e., input).
no code implementations • 28 Sep 2021 • Ketson R. M. dos Santos, Dimitrios G. Giovanis, Katiana Kontolati, Dimitrios Loukrezis, Michael D. Shields
Using this representation, geometric harmonics, an out-of-sample function extension technique, is employed to create a global map from the space of input parameters to a Grassmannian diffusion manifold.
2 code implementations • 21 Jul 2021 • Katiana Kontolati, Dimitrios Loukrezis, Ketson R. M. dos Santos, Dimitrios G. Giovanis, Michael D. Shields
For this purpose, we employ Grassmannian diffusion maps, a two-step nonlinear dimension reduction technique which allows us to reduce the dimensionality of the data and identify meaningful geometric descriptions in a parsimonious and inexpensive manner.
no code implementations • 23 Nov 2020 • Anindya Bhaduri, Christopher S. Meyer, John W. Gillespie Jr., Bazle Z. Haque, Michael D. Shields, Lori Graham-Brady
This enables the computationally feasible generation of the probabilistic velocity response (PVR) curve or the $V_0-V_{100}$ curve as a function of the impact velocity, and the ballistic limit velocity prediction as a function of the model parameters.