Understanding physics from interconnected data

TL;DR

This paper discusses the challenges of understanding complex, far-from-equilibrium physical systems like metal melting after explosions, emphasizing the need for advanced data analysis techniques to extract, reconstruct, and predict physical phenomena from noisy, multi-sensor data.

Contribution

It introduces several techniques for extracting information from noisy, entangled sensor data, laying the groundwork for future physical modeling and prediction.

Findings

Proposed methods improve data extraction from noisy images.

Techniques facilitate reconstruction of hidden physical information.

Foundation laid for predictive modeling of complex physical systems.

Abstract

Metal melting on release after explosion is a physical system far from quilibrium. A complete physical model of this system does not exist, because many interrelated effects have to be considered. General methodology needs to be developed so as to describe and understand physical phenomena involved. The high noise of the data, moving blur of images, the high degree of uncertainty due to the different types of sensors, and the information entangled and hidden inside the noisy images makes reasoning about the physical processes very difficult. Major problems include proper information extraction and the problem of reconstruction, as well as prediction of the missing data. In this paper, several techniques addressing the first problem are given, building the basis for tackling the second problem.