Theme 2
ML-Assisted Inverse Modeling

  • Watershed models involve a wide array of uncertain input parameters that are hard to observe directly;
  • Inverse modeling is challenging due to high dimensionality, computational expense;
  • Machine learning methods offer new ways to reduce dimensionality and build multi-fidelity models.