Article
Engineering, Mechanical
Mihaela Chiappetta, Chiara Piazzola, Massimo Carraturo, Lorenzo Tamellini, Alessandro Reali, Ferdinando Auricchio
Summary: The present paper aims to apply uncertainty quantification methodologies to simulate the powder bed fusion process of metal. The uncertainties of three process parameters, namely the activation temperature, the powder convection coefficient, and the gas convection coefficient, are studied in a part-scale thermomechanical model of an Inconel 625 super-alloy beam. The proposed uncertainty quantification workflow, which includes global sensitivity analysis and inverse and forward uncertainty quantification analyses, effectively reduces uncertainties and facilitates the calibration of part-scale models for powder bed fusion.
INTERNATIONAL JOURNAL OF MECHANICAL SCIENCES
(2023)
Article
Physics, Fluids & Plasmas
T. R. Younkin, K. Sargsyan, T. Casey, H. N. Najm, J. M. Canik, D. L. Green, R. P. Doerner, D. Nishijima, M. Baldwin, J. Drobny, D. Curreli, B. D. Wirth
Summary: A Bayesian inference strategy is used to estimate uncertain inputs to the global impurity transport code (GITR) modeling predictions of tungsten erosion and migration. The results are compared to experimental data to evaluate the erosion and impurity transport of tungsten in a linear plasma device.
Article
Multidisciplinary Sciences
Yikuan Li, Shishir Rao, Abdelaali Hassaine, Rema Ramakrishnan, Dexter Canoy, Gholamreza Salimi-Khorshidi, Mohammad Mamouei, Thomas Lukasiewicz, Kazem Rahimi
Summary: The study combines deep Bayesian learning with deep kernel learning to provide a more comprehensive uncertainty estimation for clinical decision making. Experimental results demonstrate that the method outperforms traditional Gaussian processes and deep Bayesian neural networks in capturing uncertainty and shows comparable generalization performance.
SCIENTIFIC REPORTS
(2021)
Article
Engineering, Multidisciplinary
Jacques H. Mclean, Matthew R. Jones, Brandon J. O'Connell, Eoghan Maguire, Tim J. Rogers
Summary: A wind turbine's power curve is vital for structural health monitoring, but existing probabilistic models often lack physical plausibility. This paper investigates two bounded Gaussian processes and demonstrates that a well-designed bounded model offers improved predictive uncertainty and accuracy compared to an unbounded model.
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2023)
Article
Nuclear Science & Technology
Riccardo Cocci, Guillaume Damblin, Alberto Ghione, Lucia Sargentini, Didier Lucor
Summary: This paper presents a methodology called Bayesian calibration for the development, validation, and uncertainty quantification of closure laws in thermal-hydraulic system codes. It introduces a robust and reliable assessment, selection, and uncertainty quantification of physical models by tuning parameters and selecting the best-suited model based on statistical indicators. The paper also discusses the application of this methodology to condensation heat transfer correlations.
ANNALS OF NUCLEAR ENERGY
(2022)
Article
Environmental Sciences
Amir H. Kohanpur, Siddharth Saksena, Sayan Dey, J. Michael Johnson, M. Sadegh Riasi, Lilit Yeghiazarian, Alexandre M. Tartakovsky
Summary: Estimating uncertainty in flood model predictions is crucial for various applications. This study focuses on uncertainty in physics-based urban flooding models, considering model complexity, uncertainty in input parameters, and the effects of rainfall intensity. The ICPR model is used to quantify floodwater depth prediction uncertainty, with results showing localized uncertainties. Model simplifications lead to overconfident predictions, while increasing model resolution reduces uncertainty but increases computational cost. The multilevel MC method is employed to reduce cost when estimating uncertainty in a high-resolution ICPR model. Utilizing ensemble estimates, the proposed framework improves flood depth forecasting accuracy compared to the ICPR model's mean prediction, even with limited measurements.
WATER RESOURCES RESEARCH
(2023)
Article
Engineering, Biomedical
Selma Metzner, Gerd Wubbeler, Sebastian Flassbeck, Constance Gatefait, Christoph Kolbitsch, Clemens Elster
Summary: Magnetic Resonance Fingerprinting (MRF) is a promising technique for fast quantitative imaging of human tissue, providing valuable diagnostic parameters like T-1 and T-2 MR relaxation times. A Bayesian approach is proposed for uncertainty quantification of dictionary-based MRF, leading to probability distributions for T-1 and T-2 in every voxel. The method successfully characterizes uncertainties in relaxation time estimates and is consistent with observed variability in simulations and in vivo measurements.
PHYSICS IN MEDICINE AND BIOLOGY
(2021)
Review
Physics, Applied
Yarin Gal, Petros Koumoutsakos, Francois Lanusse, Gilles Louppe, Costas Papadimitriou
Summary: Five researchers discuss the quantification of uncertainty in machine-learned models, focusing on issues relevant to physics problems. It is crucial to be able to measure uncertainty when comparing theoretical or computational models with observations in order to conduct sound scientific investigations. With the increasing popularity of data-driven modeling, understanding different sources of uncertainty and developing methods to estimate them has become a renewed area of interest.
NATURE REVIEWS PHYSICS
(2022)
Article
Engineering, Multidisciplinary
Panagiotis Tsilifis, Piyush Pandita, Sayan Ghosh, Valeria Andreoli, Thomas Vandeputte, Liping Wang
Summary: The text discusses the computational challenges of uncertainty propagation in complex engineering systems and the use of Gaussian Processes regression to model and quantify probability distributions of model outputs. Through introducing latent design variables and training the GP model in a lower dimensional space, the high volume of data required in high dimensional design spaces can be handled effectively.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2021)
Article
Computer Science, Artificial Intelligence
Xu Chen, Yue Zhao, Chuancai Liu
Summary: This study proposes a scalable functional variational BNN with Gaussian processes for medical image segmentation. The prior and variational posterior distributions are regarded as GPs and variational inference is performed in function space. The proposed method improves segmentation performance and uncertainty estimates compared to several state-of-the-art methods.
Article
Automation & Control Systems
Shailesh Garg, Souvik Chakraborty
Summary: VB-DeepONet is a Bayesian operator learning framework that addresses the challenges faced by the deterministic DeepONet architecture. It provides better resistance against overfitting, improved generalization, and allows for the quantification of predictive uncertainty. The results from various mechanics problems demonstrate the effectiveness of VB-DeepONet in uncertainty quantification.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Energy & Fuels
Frederik Aerts, Luca Lanzilao, Johan Meyers
Summary: This study proposes a Bayesian uncertainty quantification framework for improving wind farm models. The framework successfully distinguishes the three sources of uncertainty in the parameters and allows for model calibration and validation.
Article
Engineering, Electrical & Electronic
Paolo Manfredi
Summary: In this article, a probabilistic machine learning framework based on Gaussian process regression (GPR) and principal component analysis (PCA) is proposed for the uncertainty quantification (UQ) of microwave circuits. The proposed technique combines the inherent uncertainty of GPR pointwise predictions with the uncertainty of design parameters, providing global statistical information and confidence bounds on device performance. The application of PCA effectively deals with complex-valued output components, making it suitable for UQ of time-domain responses or multiport scattering parameters. Comparisons with polynomial chaos expansion method demonstrate that GPR achieves superior accuracy and provides prediction confidence.
IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES
(2023)
Article
Computer Science, Information Systems
Felix Fiedler, Sergio Lucia
Summary: Uncertainty quantification is essential in machine learning, but traditional neural networks struggle with this task. Bayesian neural networks address this limitation by assuming probability distributions for parameters. However, training and inference are challenging, requiring approximations. This study presents a reformulation of the log-marginal likelihood for efficient training and addresses uncertainty quantification for extrapolation points, achieving superior results compared to other methods.
Article
Public, Environmental & Occupational Health
Laura Urso, Mouhamadou Moustapha Sy, Marc-Andre Gonze, Philipp Hartmann, Martin Steiner
Summary: This study focuses on assessing conceptual model uncertainty, subtracting parameter uncertainty obtained through Bayesian inference analysis from the total uncertainty of the model output for two process-based models describing the interception of wet deposited pollutants. Quantitative evidence suggests that conceptual model uncertainty can contribute as much as, if not more than, parameter uncertainty to the total uncertainty budget of the models.
Article
Thermodynamics
Weiqi Ji, Zhuyin Ren, Youssef Marzouk, Chung K. Law
PROCEEDINGS OF THE COMBUSTION INSTITUTE
(2019)
Article
Geochemistry & Geophysics
Chen Gu, Ulrich Mok, Youssef M. Marzouk, German A. Prieto, Farrokh Sheibani, J. Brian Evans, Bradford H. Hager
GEOPHYSICAL JOURNAL INTERNATIONAL
(2020)
Article
Physics, Fluids & Plasmas
R. Sweeney, A. J. Creely, J. Doody, T. Fulop, D. T. Garnier, R. Granetz, M. Greenwald, L. Hesslow, J. Irby, V. A. Izzo, R. J. La Haye, N. C. Logan, K. Montes, C. Paz-Soldan, C. Rea, R. A. Tinguely, O. Vallhagen, J. Zhu
JOURNAL OF PLASMA PHYSICS
(2020)
Article
Physics, Fluids & Plasmas
P. Rodriguez-Fernandez, N. T. Howard, M. J. Greenwald, A. J. Creely, J. W. Hughes, J. C. Wright, C. Holland, Y. Lin, F. Sciortino
JOURNAL OF PLASMA PHYSICS
(2020)
Article
Physics, Fluids & Plasmas
A. J. Creely, M. J. Greenwald, S. B. Ballinger, D. Brunner, J. Canik, J. Doody, T. Fulop, D. T. Garnier, R. Granetz, T. K. Gray, C. Holland, N. T. Howard, J. W. Hughes, J. H. Irby, V. A. Izzo, G. J. Kramer, A. Q. Kuang, B. LaBombard, Y. Lin, B. Lipschultz, N. C. Logan, J. D. Lore, E. S. Marmar, K. Montes, R. T. Mumgaard, C. Paz-Soldan, C. Rea, M. L. Reinke, P. Rodriguez-Fernandez, K. Sarkimaki, F. Sciortino, S. D. Scott, A. Snicker, P. B. Snyder, B. N. Sorbom, R. Sweeney, R. A. Tinguely, E. A. Tolman, M. Umansky, O. Vallhagen, J. Varje, D. G. Whyte, J. C. Wright, S. J. Wukitch, J. Zhu
JOURNAL OF PLASMA PHYSICS
(2020)
Editorial Material
Physics, Fluids & Plasmas
Martin Greenwald
JOURNAL OF PLASMA PHYSICS
(2020)
Article
Physics, Fluids & Plasmas
A. Q. Kuang, S. Ballinger, D. Brunner, J. Canik, A. J. Creely, T. Gray, M. Greenwald, J. W. Hughes, J. Irby, B. LaBombard, B. Lipschultz, J. D. Lore, M. L. Reinke, J. L. Terry, M. Umansky, D. G. Whyte, S. Wukitch
JOURNAL OF PLASMA PHYSICS
(2020)
Article
Physics, Fluids & Plasmas
J. W. Hughes, N. T. Howard, P. Rodriguez-Fernandez, A. J. Creely, A. Q. Kuang, P. B. Snyder, T. M. Wilks, R. Sweeney, M. Greenwald
JOURNAL OF PLASMA PHYSICS
(2020)
Article
Computer Science, Software Engineering
Antoni Musolas, Estelle Massart, Julien M. Hendrickx, P. -A. Absil, Youssef Marzouk
Summary: The paper presents a differential geometric approach for constructing families of low-rank covariance matrices through interpolation on low-rank matrix manifolds, demonstrating its utility in practical applications such as wind field covariance approximation for unmanned aerial vehicle navigation.
BIT NUMERICAL MATHEMATICS
(2022)
Article
Physics, Fluids & Plasmas
M. Moscheni, C. Meineri, M. Wigram, C. Carati, E. De Marchi, M. Greenwald, P. Innocente, B. LaBombard, F. Subba, H. Wu, R. Zanino
Summary: As nuclear fusion experiments progress, the issue of power exhaust in future reactors remains unresolved. Different strategies are being investigated to manage power exhaust in reactor-relevant conditions, and the Divertor Tokamak Test (DTT) experiment will test these strategies. Meanwhile, the design of future tokamaks is advancing rapidly. To address the lack of reactor-relevant data, modeling is being used to reduce uncertainty in the design phase. Three state-of-the-art edge codes are compared in this study, showing encouraging agreement in certain parameters but significant disagreement in others. The results highlight limitations of the codes and potential improvements for future applications.
Article
Mathematics, Applied
Olivier Zahm, Tiangang Cui, Kody Law, Alessio Spantini, Youssef Marzouk
Summary: This paper proposes a dimension reduction technique for Bayesian inverse problems with nonlinear forward operators, non-Gaussian priors, and non-Gaussian observation noise. The likelihood function is approximated by a ridge function, and the ridge approximation is built by minimizing an upper bound on the Kullback-Leibler divergence between the posterior distribution and its approximation. The paper provides an analysis that enables control of the posterior approximation error due to sampling.
MATHEMATICS OF COMPUTATION
(2022)
Article
Physics, Fluids & Plasmas
P. Rodriguez-Fernandez, A. J. Creely, M. J. Greenwald, D. Brunner, S. B. Ballinger, C. P. Chrobak, D. T. Garnier, R. Granetz, Z. S. Hartwig, N. T. Howard, J. W. Hughes, J. H. Irby, V. A. Izzo, A. Q. Kuang, Y. Lin, E. S. Marmar, R. T. Mumgaard, C. Rea, M. L. Reinke, V Riccardo, J. E. Rice, S. D. Scott, B. N. Sorbom, J. A. Stillerman, R. Sweeney, R. A. Tinguely, D. G. Whyte, J. C. Wright, D. Yuryev
Summary: The SPARC tokamak project aims to achieve breakeven and burning plasma conditions in a compact device using high-temperature superconductor technology. It is predicted to produce high fusion power, providing important reference for the design and construction of a compact, high-field fusion power plant.
Article
Physics, Fluids & Plasmas
I. Abramovic, E. P. Alves, M. Greenwald
Summary: The prediction and control of turbulence in nuclear fusion plasmas is an important problem that is challenging to describe theoretically and model computationally. Current advanced computational models have high costs and cannot be routinely applied for plasma discharge simulation and control. The development of reduced models based on artificial neural networks shows promise, but requires extensive data and lacks extrapolation capability. In contrast, a data-driven model discovery approach based on sparse regression can infer accurate governing nonlinear partial differential equations directly from data, and can be extrapolated to unexplored parameter ranges.
JOURNAL OF PLASMA PHYSICS
(2022)
Article
Physics, Fluids & Plasmas
A. J. Creely, D. Brunner, R. T. Mumgaard, M. L. Reinke, M. Segal, B. N. Sorbom, M. J. Greenwald
Summary: The SPARC tokamak's unique capabilities have the potential to significantly contribute to tokamak science and plasma physics, encouraging collaboration and broader data access beyond the CFS and MIT teams.
PHYSICS OF PLASMAS
(2023)
Article
Mathematics, Applied
Johnathan M. Bardsley, Tiangang Cui, Youssef M. Marzouk, Zheng Wang
SIAM JOURNAL ON SCIENTIFIC COMPUTING
(2020)