Scientific machine learning is at the core of modern computational technology; it has the power to potentially transform research in science and engineering. While machine learning methods have been much used with success, there are still tremendous challenges and opportunities for increasing the scale, rigor, robustness, and reliability of such methods for practical applications.
The goal of this Research Topic is to collect recent state-of-the-art research in advancing the fundamental understanding of scientific machine learning, as well as its applications in enhancing modeling and simulation. Expected contributions include, but are not limited to:
1) leveraging domain knowledge, such as physical principles and invariance structures, to machine learning models to improve its accuracy and defensibility, as well as accelerating the model training;
2) incorporating statistics, uncertainty quantification, and probabilistic modeling into scientific machine learning to deal with large scale complex models and data;
3) improving upon current numerical solvers through the judicious use of machine learning algorithms and developing new machinery to optimally manage the interplay between traditional and machine learning models.
Scientific machine learning is at the core of modern computational technology; it has the power to potentially transform research in science and engineering. While machine learning methods have been much used with success, there are still tremendous challenges and opportunities for increasing the scale, rigor, robustness, and reliability of such methods for practical applications.
The goal of this Research Topic is to collect recent state-of-the-art research in advancing the fundamental understanding of scientific machine learning, as well as its applications in enhancing modeling and simulation. Expected contributions include, but are not limited to:
1) leveraging domain knowledge, such as physical principles and invariance structures, to machine learning models to improve its accuracy and defensibility, as well as accelerating the model training;
2) incorporating statistics, uncertainty quantification, and probabilistic modeling into scientific machine learning to deal with large scale complex models and data;
3) improving upon current numerical solvers through the judicious use of machine learning algorithms and developing new machinery to optimally manage the interplay between traditional and machine learning models.