Accelerating SED Modeling of Astrophysical Objects Using Neural Networks

TL;DR

The paper addresses the bottleneck of fitting blazar SEDs in lepto-hadronic models by training a GRU-based neural network surrogate on large simulations from AM3 and LeHaMoC to rapidly predict the photon spectrum. The surrogate enables efficient Bayesian inference, achieving close agreement with full numerical codes (differences up to about 20%) and substantial speedups (hours instead of thousands of hours). A blind MCMC fitness test shows recovered posteriors are generally near the true parameters, with degeneracies tied to energy-band coverage. The work provides an open-access, scalable tool for fast SED analysis and sets the stage for future extensions including neutrino spectra.

Abstract

Interpreting the spectral energy distributions (SEDs) of astrophysical objects with physically motivated models is computationally expensive. These models require solving coupled differential equations in high-dimensional parameter spaces, making traditional fitting techniques such as Markov Chain Monte Carlo or nested sampling prohibitive. A key example is modeling non-thermal emission from blazar jets - relativistic outflows from supermassive black holes in Active Galactic Nuclei that are among the most powerful emitters in the Universe. To address this challenge, we employ machine learning to accelerate SED evaluations, enabling efficient Bayesian inference. We generate a large sample of lepto-hadronic blazar emission models and train a neural network (NN) to predict the photon spectrum with strongly reduced run time while preserving accuracy. As a proof of concept, we present an NN-based tool for blazar SED modeling, laying the groundwork for future extensions and for providing an open-access resource for the astrophysics community.

Figures (1)

• Figure 1: Corner plot of the posterior distribution of model parameters computed by the NN fit to the simulated data. In the upper inlay panel we display the SED with the simulated dataset, model estimated based on numerical codes and 100 predictions of the NN model.