Data-driven multiscale modeling for correcting dynamical systems
This method addresses challenges in dynamical systems prediction, particularly for climate modeling, but appears incremental as it builds on existing multiscale and correction techniques.
The authors tackled the problem of predicting quantities in dynamical systems by proposing a multiscale approach that corrects chaotic underlying models, specifically applied to a climate subgrid parameterization task to reflect unresolved fine-scale dynamics.
We propose a multiscale approach for predicting quantities in dynamical systems which is explicitly structured to extract information in both fine-to-coarse and coarse-to-fine directions. We envision this method being generally applicable to problems with significant self-similarity or in which the prediction task is challenging and where stability of a learned model's impact on the target dynamical system is important. We evaluate our approach on a climate subgrid parameterization task in which our multiscale networks correct chaotic underlying models to reflect the contributions of unresolved, fine-scale dynamics.