Abstract
The development of physics-informed deep learning is radically changing compu-tational science and engineering, allowing for an effective integration ofphysics-based and datadriven modeling. Deep learning provides a powerful tool forthe
discovery of governing dynamics underneath data and enables nonlinear model-reduction. A Bayesian viewpoint of deep learning is discussed in this chapter towards the quantification of modeling uncertainties in physics-informed deep learning.
discovery of governing dynamics underneath data and enables nonlinear model-reduction. A Bayesian viewpoint of deep learning is discussed in this chapter towards the quantification of modeling uncertainties in physics-informed deep learning.
| Original language | English |
|---|---|
| Title of host publication | Mathematics: Key Enabling Technology for Scientific Machine Learning |
| Editors | W. H. A. Schilders |
| Publisher | Dutch Platform for Mathematics |
| Pages | 47-51 |
| Publication status | Published - 2021 |
| Externally published | Yes |
Subject classification (UKÄ)
- Other Computer and Information Science