Gaussian Process Interpolations for AMR & Super Resolution

Adaptive Mesh Refinements (AMR) & Super Resolution

Research Group: S. Reeves, D. Lee (UCSC)

We develop a new interpolation approach using Gaussian Process Modeling. The approach has been extended to produce high-order accurate interpolation method for configuring computational grids adaptively called adaptive mesh refinements in massively parallel simulations.

Another area of application is to reconstruct high-resolution images based on convolutional neural network (CNN).