Dust Neural nEtworks Technology (DustNET) for multiscale plasma physics*

Ubiquitous in the observable universe, dusty plasmas provide fertile ground to studying and testing multiscale plasma physics, as well as to developing technologies of broad interest, including, for example, microelectronics and carbon-neutral energy. Advances in high-resolution measurement of dust grains [1], neural network algorithms, and a growing number of dusty plasma image datasets in diverse settings motivate Dust Neural nEtworks Technology (DustNET), mirroring the well-known ImageNet, which has been instrumental in advancing computer vision and deep learning research. Here we describe examples of existing datasets and neural network architectures, which form the initial building blocks of DustNET.  Data fusion of experimental data, numerical data, and other synthetic and meta data, such as ones from generative artificial intelligence (AI), offer additional options for DustNET construction. Besides applications in experimental data processing, data interpretation, predictions (inferences) and uncertainty quantification, DustNET-enabled deep neural networks may also be used for real-time experimental dusty plasma controls and optimization [2], and searching for new physics beyond the Standard Model of physics. This work is supported in part by the DoE Fusion Energy Sciences. LANL release number LA-UR- 24-25332.

 

[1] Zhehui Wang, Jiayi Xu, Yao E Kovach, Bradley T Wolfe, Edward Thomas, Hanqi Guo, John E Foster, Han-Wei Shen, Phys. Plasma. 27 (2020) 033703.

 

[2] Zhehui Wang, Shanny Lin, Miles Teng-Levy, Pinghan Chu, Bradley T Wolfe, Chun-Shang Wong, Christopher S Campbell, Xin Yue, Liyuan Zhang, Derek Aberle, Mariana Alvarado Alvarez, David Broughton, Ray T Chen, Baolian Cheng, Feng Chu, Eric R Fossum, Mark A Foster, Chengkun Huang, Velat Kilic, Karl Krushelnick, Wenting Li, Eric Loomis, Thomas Schmidt Jr, Sky K Sjue, Chris Tomkins, Dmitry A Yarotski, Renyuan Zhu, arXiv preprint arXiv:2401.08390 (2024); https://arxiv.org/abs/2401.08390.