In this study, the fundamental stellar atmospheric parameters (Teff, log g, [Fe/H] and [{\alpha}/Fe]) were derived for low-resolution spectroscopy from LAMOST DR5 with Generative Spectrum Networks (GSN). This follows the same scheme as a normal artificial neural network with stellar parameters as the input and spectra as the output. The GSN model was effective in producing synthetic spectra after training on the PHOENIX theoretical spectra. In combination with Bayes framework, the application for analysis of LAMOST observed spectra exhibited improved efficiency on the distributed computing platform, Spark. In addition, the results were examined and validated by a comparison with reference parameters from high-resolution surveys and asteroseismic results. Our results show good consistency with the results from other survey and catalogs. Our proposed method is reliable with a precision of 80 K for Teff, 0.14 dex for log g, 0.07 dex for [Fe/H] and 0.168 dex for [{\alpha}/Fe], for spectra with a signal-to-noise in g bands (SNRg) higher than 50. The parameters estimated as a part of this work are available at http://paperdata.china-vo.org/GSN_parameters/GSN_parameters.csv.
@article{arxiv.1811.08207,
title = {Analysis of Stellar Spectra from LAMOST DR5 with Generative Spectrum Networks},
author = {Wang Rui and Luo A-li and Zhang Shuo and Hou Wen and Du Bing and Song Yi-Han and Wu Ke-Fei and Chen Jian-Jun and Zuo Fang and Qin Li and Chen Xiang-Lei and Lu Yan},
journal= {arXiv preprint arXiv:1811.08207},
year = {2019}
}