Transformer model for evaluating Gaussian beam propagation through turbulent atmosphere

Xun Cui; Xiaowei Chen; Xianmei Qian; Wenyue Zhu; Pengfei Wu; Lingyun Min; Ruizhong Rao

doi:10.1117/12.3017798

30 April 2024 Transformer model for evaluating Gaussian beam propagation through turbulent atmosphere

Xun Cui, Xiaowei Chen, Xianmei Qian, Wenyue Zhu, Pengfei Wu, Lingyun Min, Ruizhong Rao

Author Affiliations +

Proceedings Volume 13157, Sixth Conference on Frontiers in Optical Imaging and Technology: Applications of Imaging Technologies; 131570T (2024) https://doi.org/10.1117/12.3017798
Event: Sixth Conference on Frontiers in Optical Imaging Technology and Applications (FOI2023), 2023, Nanjing, JS, China

Abstract

Scaling law and artificial intelligence model are two methods to quickly evaluate the far-field spot characteristics of laser propagation through turbulence. On the one hand, it is necessary to compare the evaluation accuracy of both models. On the other hand, the comparison between different models is only meaningful if each has their best accuracy. For specific scaling law model and artificial intelligence model, scaling exponents and hyperparameters determine the evaluation accuracy of the model to a certain extent. This paper discusses how to search better scaling exponents and hyperparameters to construct each model and compare the evaluation accuracy of both models. This paper first introduces the MRSS (Modified-Radius-Square-Sum) scaling law model and FT-Transformer (Feature Tokenizer + Transformer) model, and 3 hyperparameter (scaling exponent) optimization algorithms. Then, the accuracy of scaling exponents and hyperparameters obtained by different optimization algorithms is compared. Finally, the best scaling exponents and hyperparameters are used to construct each model. The results show that the TPE algorithm achieves better search results in fewer iterations for the FT-Transformer model, and the CmaEs algorithm achieves higher accuracy in more iterations for the scaling law model. The FT-Transformer model has better accuracy compared to the scaling law model, with the mean relative error of the far field effective radius and mean intensity is 1.32% and 2.66%, while those of scaling law model is 1.97% and 3.91% respectively.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Xun Cui, Xiaowei Chen, Xianmei Qian, Wenyue Zhu, Pengfei Wu, Lingyun Min, and Ruizhong Rao "Transformer model for evaluating Gaussian beam propagation through turbulent atmosphere", Proc. SPIE 13157, Sixth Conference on Frontiers in Optical Imaging and Technology: Applications of Imaging Technologies, 131570T (30 April 2024); https://doi.org/10.1117/12.3017798

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

;

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE