Computer Methods in Applied Mechanics and Engineering
Advanced Machine Learning for Uncertainty Quantification
摘要截稿:
全文截稿: 2024-06-30
影响因子: 5.763
期刊难度:
CCF分类: 无
中科院JCR分区:
• 大类 : 工程技术 - 1区
• 小类 : 工程:综合 - 1区
• 小类 : 数学跨学科应用 - 1区
• 小类 : 力学 - 1区
Overview
In the field of Computational Science and Engineering, a disruptive force has emerged over recent years, reshaping the boundaries of what we can achieve: Machine Learning. With its ability to extract patterns and insights from vast amounts of data, machine learning has pushed computational science and engineering to new heights, unlocking innovative solutions to intricate problems and paving the way for a new era in predictive science.
Given the increasing complexity of engineered systems, it has become imperative to create cost-effective predictive models for various challenging applications such as reliability analysis, uncertainty quantification, and system design optimization. Machine Learning has demonstrated its potential in creating data-driven surrogate models that can replace the computationally expensive physics-based high-fidelity models. These surrogate models not only help with uncertainty quantification but also facilitate the continuous improvement of system design in the presence of uncertainties.
Despite these advancements, there are ongoing difficulties in enhancing the accuracy of predictive models that simulate stochastic engineering systems. Focusing on the intersection of machine learning and uncertainty modeling, the issue aims to bring together contributions that push the boundaries of current practices. The goal is to provide a comprehensive overview of the state-of-the-art in advanced machine learning techniques dedicated to quantifying uncertainties, fostering a deeper understanding of uncertainty in complex systems, and promoting the development of more robust and reliable models.
Topic Areas
Surrogate models for forward and inverse UQ
Multi-fidelity and multi-level UQ using machine learning
Statistical surrogate models
Probabilistic manifold learning
Data-driven material science
Deep Bayesian models for Uncertainty Propagation
Sparse Gaussian Processes for Efficient UQ
Time Series Uncertainty Modeling
Machine learning-enhanced iterative solvers for large-scale stochastic problems
Linear and nonlinear dimensionality reduction techniques
Uncertainty quantification of machine learning models
Applications of advanced machine learning methods for uncertainty quantification in Computational Fluid Dynamics, Structural Analysis, Multiscale and Multiphysics problems
Guest editors:
Prof. Vissarion PapadopoulosAffiliation: National Technical University of Athens (NTUA), Athens, Greece
Prof. Eleni ChatziAffiliation: ETH Zürich, Zurich, Switzerland
Prof. Roger GhanemAffiliation: University of Southern California, Los Angeles, United States
Prof. Christian SoizeAffiliation: Université Gustave Eiffel, Champs-sur-Marne, France