Learning the Relationship Between Galaxies's Spectra and Their Star Formation Histories Using Convolutional Neural Networks and Cosmological Simulations

Christopher Lovell (U. Hertfordshire)

I will present a novel method for inferring galaxy star formation histories (SFH) using machine learning methods coupled with two cosmological hydrodynamic simulations. The method uses convolutional neural networks to learn the relationship between synthetic galaxy spectra and high-resolution SFH's from the EAGLE and Illustris models. We achieve high test accuracy on both dust attenuated and intrinsic spectra, including the effects of observational noise, and make estimates for the observational and modelling errors. To further evaluate the generalization properties we apply models trained on one simulation to spectra from the other, which leads to only a small increase in the error. We apply each trained model to SDSS DR7 spectra, and find smoother histories than in the VESPA catalogue. This new approach complements the results of existing spectral-energy-distribution fitting techniques, providing SFHs directly motivated by the results of the latest cosmological simulations, as well as opportunities for using the outputs of such models within existing bayesian SED fitting pipelines.