Extracting Forces from Noisy Dynamics in Dusty Plasmas

TL;DR

This paper demonstrates that supervised machine learning can accurately extract multiple environmental forces acting on particles in dusty plasmas from noisy data, outperforming traditional methods and enabling non-contact measurements of key parameters.

Contribution

The study introduces a machine learning approach trained on simulated data to accurately infer forces and parameters in dusty plasmas, surpassing conventional techniques.

Findings

50% improvement in parameter prediction accuracy

Non-contact measurement of particle charge and Debye length

Effective extraction of electrostatic, hydrodynamic, and stochastic forces

Abstract

Extracting environmental forces from noisy data is a common yet challenging task in complex physical systems. Machine learning represents a robust approach to this problem, yet is mostly tested on simulated data with known parameters. Here we use supervised machine learning to extract the electrostatic, hydrodynamic, and stochastic forces acting on micron-sized charged particles levitated in an argon plasma. Trained on simulated particle trajectories using more than 100 dynamical and statistical features, the model predicts system parameters with 50\% better accuracy than conventional methods, and provides non-contact measurements of the particle charge and Debye length.