Recent developments in machine learning techniques have been supported by the continuous increase in availability of high-performance computational resources and data. While large volumes of data are needed to train machine learning models, the community is looking for improved training techniques to optimize and speed up the process. One of the effective ways towards resolving this problem, Physics-Informed Machine Learning, suggests using prior available knowledge in the form of physical laws and equations to improve the training of machine learning models, making it more efficient, and resulting in robust and trustworthy predictive models.
In this Collection, we aim to bring together research of Theoretical and computational frameworks, Data-driven predictive models, Data-driven scientific discovery in physics and engineering, and invite Commentary from experts.
Topics of interest include but are not limited to the following:
- Theoretical foundations of physics-informed machine learning
- Physics-informed neural networks and their modifications
- Physics-informed deep learning
- Equivariant neural networks
- Neural operators
- Data-driven predictive models
- Predictive models based on limited or damaged data
- Data-driven scientific discovery in physics and engineering
- Application in computational fluid dynamics
Leave a comment