This project includes data Jupyter notebooks for analysis of performance of Active Learning techniques applied to Adaptive Mesh Refinement (AMR) simulations. Specifially, we consider AMR simulations of a shock-bubble interaction phenomenon. We train and evaluate Active Learning algorithms on the data from AMR simulation runs performed on Edison supercomputer at NERSC (US National Energy Research Computing Center).
Contributors:
- Dmitry Duplyakin [email protected]
- Jed Brown [email protected]
- Donna Calhoun [email protected]