Thermodynamics-informed super-resolution of scarce temporal dynamics data

arXiv:2402.17506v1 Announce Type: cross
Abstract: We present a method to increase the resolution of measurements of a physical system and subsequently predict its time evolution using thermodynamics-aware neural networks. Our method uses adversarial autoencoders, which reduce the dimensionality of the full order model to a set of latent variables that are enforced to match a prior, for example a normal distribution. Adversarial autoencoders are seen as generative models, and they can be trained to generate high-resolution samples from low-resoution inputs, …

abstract adversarial arxiv autoencoders cs.lg data dimensionality dynamics evolution its time networks neural networks physics.comp-ph reduce set temporal type variables