Article
Engineering, Mechanical
P. L. Green, L. J. Devlin, R. E. Moore, R. J. Jackson, J. Li, S. Maskell
Summary: This paper discusses the optimization of the 'L-kernel' in Sequential Monte Carlo samplers to improve performance, resulting in reduced variance of estimates and fewer resampling requirements.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Statistics & Probability
Lewis J. Rendell, Adam M. Johansen, Anthony Lee, Nick Whiteley
Summary: In order to conduct Bayesian inference with large datasets, it is beneficial to distribute the data across multiple machines. By introducing an instrumental hierarchical model and using an SMC sampler with a sequence of association strengths, approximations of posterior expectations can be improved and the association strength can be adjusted accordingly.
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS
(2021)
Article
Engineering, Mechanical
Adolphus Lye, Alice Cicirello, Edoardo Patelli
Summary: This tutorial paper reviews the use of advanced Monte Carlo sampling methods in Bayesian model updating for engineering applications, introducing different methods and comparing their performance. Three case studies demonstrate the advantages and limitations of these sampling techniques in parameter identification, posterior distribution sampling, and stochastic identification of model parameters.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2021)
Review
Statistics & Probability
Christopher Nemeth, Paul Fearnhead
Summary: MCMC algorithms are considered the gold standard technique for Bayesian inference, but the computational cost can be prohibitive for large datasets, leading to the development of scalable Monte Carlo algorithms. One type of these algorithms is SGMCMC, which reduces per-iteration cost by utilizing data subsampling techniques.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2021)
Article
Mathematics
Samuel Livingstone
Summary: This study investigates the impact of proposal distributions on the ergodicity of the Metropolis-Hastings method, showing that suitable choices can alter the ergodic properties of the algorithm. It is found that allowing the proposal variance to grow unboundedly in the tails of heavy-tailed distributions can establish geometric ergodicity, but the growth rate needs to be carefully controlled to avoid high rejection rates. Furthermore, a judicious choice of proposal distribution can lead to geometric ergodicity in scenarios where probability concentrates on narrower tails, which is not the case for the Random Walk Metropolis.
Article
Engineering, Environmental
Babak Jamhiri, Yongfu Xu, Mahdi Shadabfar, Susanga Costa
Summary: A probabilistic machine learning framework is developed to improve deterministic models, utilizing a complete set of data-driven soil and environment parameters as inputs to predict crack surface ratio. Monte Carlo simulation is employed to insert uncertainties in the models, and two sensitivity analyses are conducted to assess prediction reliability. Results show that GBTs perform the best in terms of prediction accuracy and parameter importance analysis ranks FWC as the most important factor.
BULLETIN OF ENGINEERING GEOLOGY AND THE ENVIRONMENT
(2023)
Article
Ecology
Luiza Guimaraes Fabreti, Sebastian Hoehna
Summary: This study explores different methods for assessing convergence in phylogenetics, including deriving a threshold for minimum effective sample size and converting tree samples into traces of absence/presence of splits for standard ESS computation. The Kolmogorov-Smirnov test is suggested for assessing convergence in distribution between replicated MCMC runs, while potential scale reduction factor is deemed biased for skewed posterior distributions. Additionally, the study introduces a method for computing distribution of differences in split frequencies, highlighting the importance of using the 95% quantile for checking convergence in split frequencies.
METHODS IN ECOLOGY AND EVOLUTION
(2022)
Article
Engineering, Civil
Jia-Hua Yang, Heung-Fai Lam, Yong-Hui An
Summary: The paper proposes a new two-phase adaptive MCMC method to address the problem of determining the posterior probability density function (PDF) in Bayesian model updating. By using a parameter-space search algorithm and a weighted MCMC algorithm, samples in the regions of high probability can be generated adaptively without going through computationally demanding multiple levels.
ENGINEERING STRUCTURES
(2022)
Article
Automation & Control Systems
Maxime Vono, Daniel Paulin, Arnaud Doucet
Summary: This paper investigates the computational challenges of exact Bayesian inference for complex models and proposes a split Gibbs sampler algorithm as an alternative approach. The theoretical analysis, supported by numerical illustrations, suggests that this algorithm performs well in high-dimensional scenarios.
JOURNAL OF MACHINE LEARNING RESEARCH
(2022)
Article
Engineering, Manufacturing
P. Honarmandi, R. Seede, L. Xue, D. Shoukr, P. Morcos, B. Zhang, C. Zhang, A. Elwany, I. Karaman, R. Arroyave
Summary: The Eagar-Tsai (E-T) model in the context of 3D printing was studied systematically from an uncertainty quantification/propagation (UQ/UP) perspective. Model parameters were calibrated against experimental data using Markov Chain Monte Carlo (MCMC) sampling, and posterior distributions of parameter values were propagated. It was found that discrepancies between predicted and measured melt pool depths existed under keyholing conditions, but a physics-based correction improved agreement with experiments without increasing model complexity significantly.
ADDITIVE MANUFACTURING
(2021)
Article
Mathematics, Applied
Nikolaj T. Mucke, Benjamin Sanderse, Sander M. Bohte, Cornelis W. Oosterlee
Summary: In the context of solving inverse problems in physics using Bayesian inference, a new approach called Markov Chain Generative Adversarial Neural Network (MCGAN) is proposed to reduce computational costs. By training a GAN to sample from a low-dimensional latent space and incorporating it into a Markov Chain Monte Carlo method, efficient sampling from the posterior distribution is achieved, replacing the need for high-dimensional priors and expensive forward mappings. The proposed methodology converges to the true posterior in Wasserstein-1 distance and sampling from the latent space is weakly equivalent to sampling in the high-dimensional space.
COMPUTERS & MATHEMATICS WITH APPLICATIONS
(2023)
Article
Computer Science, Interdisciplinary Applications
Riko Kelter
Summary: This paper introduces a R package that performs Bayesian inference in ANOVA, focusing on effect size estimation instead of hypothesis testing with full posterior inference implemented via MCMC.
Article
Computer Science, Theory & Methods
Alix Marie d'Avigneau, Sumeetpal S. Singh, Lawrence M. Murray
Summary: Efficient MCMC algorithms are crucial in Bayesian inference, especially in the context of parallel tempering. This study addresses the issue of randomly varying local move completion times in multi-processor parallel tempering by imposing real-time deadlines on the parallel local moves and performing exchanges at these deadlines without any processor idling. The methodology of exchanges at real-time deadlines is shown to lead to significant performance enhancements without introducing bias, with potential applications in ABC algorithms for parameter estimation.
STATISTICS AND COMPUTING
(2021)
Article
Mechanics
Lingyu Yue, Marie-Claude Heuzey, Jonathan Jalbert, Martin Levesque
Summary: A framework based on Bayesian inference is proposed in this study to identify the minimum parameter set in linear viscoelastic constitutive theories. Experimental validation demonstrates the robustness and adequacy of this method.
INTERNATIONAL JOURNAL OF SOLIDS AND STRUCTURES
(2021)
Article
Engineering, Industrial
Zeyu Wang, Abdollah Shafieezadeh
Summary: This paper presents a new approach to overcome the computational cost problem of Bayesian updating for complex computational models. It decomposes the updating problem into a set of sub-reliability problems with uncertain failure thresholds, enabling precise identification of intermediate failure thresholds and training of surrogate models. The proposed method reduces computational costs significantly while maintaining high accuracy.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Environmental Sciences
Ishfaq Rashid Sheikh, K. M. N. Saquib Wani, Fazal E. Jalal, Mohammad Yousuf Shah
Summary: The strength and rigidity of the base course are crucial for pavement performance. The use of geosynthetic materials to reinforce quarry waste can improve its load-bearing capacity. Artificial neural network analysis can predict deformation on the pavement.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2022)
Article
Chemistry, Physical
Hanif Ullah, Mudassir Iqbal, Kaffayatullah Khan, Arshad Jamal, Adnan Nawaz, Nayab Khan, Fazal E. Jalal, Abdulrazak H. Almaliki, Enas E. Hussein
Summary: This research investigates the performance of rubberized concrete by using waste tires as a replacement for fine aggregates. The stress-strain behavior of the rubberized concrete is compared with established analytical models. The study suggests potential applications of rubberized concrete as a structural material.
Article
Engineering, Marine
Mudassir Iqbal, Daxu Zhang, Fazal E. Jalal
Summary: This study develops a random forest regression model to predict the tensile strength retention (TSR) of laboratory conditioned GFRP bars in alkaline environment. Sensitivity and parametric analysis show that temperature, pH, volume fraction of fibers, conditioning duration, and bar diameter are influential attributes in TSR. The existing recommendations by various structural codes regarding environmental reduction factors are conservative and need revision.
JOURNAL OF OCEAN ENGINEERING AND SCIENCE
(2022)
Article
Engineering, Civil
Kennedy C. Onyelowe, Fazal E. Jalal, Mudassir Iqbal, Zia Ur Rehman, Kizito Ibe
Summary: Gene expression programming (GEP) and multi-expression programming (MEP) were used to predict the unconfined compressive strength of soil under unsaturated conditions. GEP outperformed MEP and multiple linear regression models, showing high accuracy and minimal errors in training and validation. Sensitivity analysis indicated the parameters influencing the model's performance.
INNOVATIVE INFRASTRUCTURE SOLUTIONS
(2022)
Article
Engineering, Civil
Babak Jamhiri, Yongfu Xu, Fazal E. Jalal, Yang Chen
Summary: This study investigates the relationship between undrained shear strength and B-ratio, void ratio, confinement pressure, and principal stress difference in zeolite-lime-treated fine sands through comprehensive experimental research. Based on the experimental evidence, a novel trend-adjusted (TA) growth forecast is performed to extend the curing ages beyond the experimental program conditions. Furthermore, a hybrid artificial neural network (ANN) model is proposed, which shows improved optimization and high accuracy in predicting undrained shear resistance considering extended curing periods. Results of variable importance and sensitivity analysis highlight the significant impact of underlying degree of saturation on shear resistance, followed by void ratio, confinement pressure, and zeolite content.
TRANSPORTATION INFRASTRUCTURE GEOTECHNOLOGY
(2023)
Article
Engineering, Geological
Guosheng Xiang, Weimin Ye, Fazal E. Jalal, Zhijie Hu
Summary: Investigation of the effect of saline solutions on the shear strength of compacted bentonite revealed that the concentration of the saline solution has a significant impact on the peak stress and the ordering of the shear strength. Cation exchange process plays a major role in affecting the shear strength of bentonite. The angle of internal friction is minimally affected by the solution concentration.
ENGINEERING GEOLOGY
(2022)
Article
Green & Sustainable Science & Technology
Aminul Haque, Bing Chen, Muhammad Faisal Javed, Fazal E. Jalal
Summary: This study uses AI models to predict the mechanical strength of MPC-FA compounds, and the results show that the DNN2 and OGPR methods have high prediction accuracy. Sensitivity analysis reveals that the FA content has the main impact on the strength of MPC-FA mixtures. These predictions can be applied in practical fields to reduce workload, labor, and material consumption through optimizing mix combinations.
JOURNAL OF CLEANER PRODUCTION
(2022)
Article
Chemistry, Physical
Kaffayatullah Khan, Babatunde Abiodun Salami, Mudassir Iqbal, Muhammad Nasir Amin, Fahim Ahmed, Fazal E. Jalal
Summary: This study utilizes artificial intelligence models to analyze the optimal ratio of ground-granulated blast furnace slag (GGBFS) and fly ash (FA) to the binder content. The gradient boosting tree (GBT) model is found to have the highest accuracy. Sensitivity analysis reveals that aging of the concrete is the most influential parameter.
Article
Engineering, Marine
Mudassir Iqbal, Khalid Elbaz, Daxu Zhang, Lili Hu, Fazal E. Jalal
Summary: This study used particle swarm optimization, genetic algorithm, and support vector machine to optimize the adaptive neuro-fuzzy inference system model for more accurate prediction of the tensile strength of GFRP bars in alkaline environments. Through the collection of experimental samples and k-fold cross-validation, robust and reliable prediction models were developed.
JOURNAL OF OCEAN ENGINEERING AND SCIENCE
(2023)
Article
Environmental Sciences
Babak Jamhiri, Mahdi Shadabfar, Fazal E. Jalal
Summary: Periodic cycles of flood and drought aggravated by global warming induce critical desiccation cracks in soils. This study presents an alternative framework to assess the probability of crack propagation using Monte Carlo sampling and Gaussian random fields. The results show that matric suction plays a governing role in crack propagation, and crack propagation tends to decrease with increased tensile strength and reduced matric suction. The probability of crack propagation is directly related to soil compaction density reduction and variations of matric suction. Random field sampling is superior to MCS in estimating crack propagation. Considering spatial uncertainty in measuring crack propagation results in dependable estimations with only 8% deviation from field observation. The developed probabilistic framework provides a promising alternative for reliable design without demanding experiments and complex simulations.
MODELING EARTH SYSTEMS AND ENVIRONMENT
(2023)
Article
Energy & Fuels
Babak Jamhiri, Yongfu Xu, Mahdi Shadabfar, Fazal E. Jalal
Summary: The prediction of thermal crack propagation in desiccated soils is imperfect. To address this issue, a probabilistic framework is developed to enhance the crack estimation reliability. The results show that cracking probability is imminent in near-surface layers.
GEOMECHANICS FOR ENERGY AND THE ENVIRONMENT
(2023)
Article
Engineering, Environmental
Babak Jamhiri, Yongfu Xu, Mahdi Shadabfar, Susanga Costa
Summary: A probabilistic machine learning framework is developed to improve deterministic models, utilizing a complete set of data-driven soil and environment parameters as inputs to predict crack surface ratio. Monte Carlo simulation is employed to insert uncertainties in the models, and two sensitivity analyses are conducted to assess prediction reliability. Results show that GBTs perform the best in terms of prediction accuracy and parameter importance analysis ranks FWC as the most important factor.
BULLETIN OF ENGINEERING GEOLOGY AND THE ENVIRONMENT
(2023)
Article
Engineering, Multidisciplinary
Babak Jamhiri, Fazal E. Jalal, Yang Chen
Summary: This study conducted a comprehensive experimental program to investigate the relationships between B-ratio, void ratio, and principal stress difference at failure of zeolite-alkali activated sands. The proposed unified relationships presented by this research provide a solid framework for the treatment of fine sands with natural zeolite-lime blends.
MULTISCALE AND MULTIDISCIPLINARY MODELING EXPERIMENTS AND DESIGN
(2022)