Turbulent flow modeling remains a significant challenge due to the computational expense and difficulty in accurately capturing fine-scale structures. This work introduces an adversarially trained neural operator that effectively combines operator learning with generative modeling, achieving substantial improvements in spatio-temporal super resolution, forecasting, and sparse flow reconstruction, thereby enabling accurate and efficient analysis of turbulent flows.
- Vivek Oommen
- Siavash Khodakarami
- George Em Karniadakis
