AGN X-ray spectroscopy with neural networks

ABSTRACT We explore the possibility of using machine learning to estimate physical parameters directly from active galactic nucleus (AGN) X-ray spectra without needing computationally expensive spectral fitting. Specifically, we consider survey quality data, rather than long pointed observations, ensure that this approach works in regime where it is most likely be applied. simulate Athena Wide Field Imager AGN with warm absorbers, and train simple neural networks ionization column density absorbers. find can give comparable accuracy fitting, risk outliers caused by fit sticking a false minimum, an improvement around three orders magnitude speed. also demonstrate principal component analysis reduce dimensionality data prior inputting into net significantly increase parameter estimation for negligible computational cost, while allowing simpler network architecture used.