Core Concepts
Novel dimension reduction technique for exact reconstruction in phase space.
Abstract
The article introduces a novel dimension reduction technique to enable exact reconstruction of time-dependent data in phase space. By projecting phase space onto lower-dimensional subspaces, the curse of high dimensionality is circumvented. The study focuses on superresolution, proving that known exact reconstruction results hold after dimension reduction. Additionally, new error estimates for reconstructions from noisy data are provided using optimal transport metrics. The proposed method is applicable beyond superresolution scenarios.
Stats
Known exact reconstruction results stay true after dimension reduction.
New error estimates of reconstructions from noisy data are provided.