Deep learning model for multiwavelength emission from low-luminosity active galactic nuclei

TL;DR

This paper introduces a machine learning model that efficiently generates and fits multiwavelength spectral energy distributions of low-luminosity active galactic nuclei, significantly reducing computational costs compared to traditional methods.

Contribution

The study presents a neural network-based approach to model LLAGN SEDs, enabling rapid and accurate fitting of observational data with much less computational effort.

Findings

ML model reproduces radio-to-X-ray emission with high precision

Model fits LLAGN SEDs of M87, NGC 315, NGC 4261 successfully

Fitting process is 400,000 times faster than previous models

Abstract

Most active supermassive black holes (SMBH) in present-day galaxies are underfed and consist of low-luminosity active galactic nuclei (LLAGN). They have multiwavelength broadband spectral energy distributions (SED) dominated by non-thermal processes which are quite different from those of the brighter, more distant quasars. Modelling the observed SEDs of LLAGNs is currently challenging, given the large computational expenses required. In this work, we used machine learning (ML) methods to generate model SEDs and fit sparse observations of LLAGNs. Our ML model consisted of a neural network and reproduced with excellent precision the radio-to-X-rays emission from a radiatively inefficient accretion flow around a SMBH and a relativistic jet, at a small fraction of the computational cost. The ML method performs the fit $4 \times 10^5$ times faster than previous semianalytic models. As a proof-of-concept, we used the ML model to reproduce the SEDs of the LLAGNs M87, NGC 315 and NGC 4261.