Beyond the Brightest: A Deep Learning Approach to Identifying Major and Minor Galaxy Mergers in CANDELS at z ∼ 1

Galaxy mergers play an important role in galaxy evolution. Therefore, accurate merger identifications are paramount for achieving a complete understanding of how galaxies evolve. As we enter the era of large, deep, high-resolution imaging surveys, we can observe mergers extending to even lower masses and higher redshifts. Despite low-mass galaxies being more common, many previous merger identification methods were calibrated for high-mass galaxies, which are easier to identify. To prepare for upcoming surveys, we train a convolutional neural network (CNN) using mock Hubble Space Telescope Cosmic Assembly Near-IR Deep Extragalactic Legacy Survey (CANDELS) images at z ∼ 1 created from the IllustrisTNG50 cosmological simulation. We successfully identify galaxy mergers between a wide range of galaxies (108M⊙ < M⋆ < 1012.5M⊙, and q ≥ 1: 10), achieving an overall accuracy, purity, and completeness of ∼65%. We show, for the first time, that a CNN trained on this diverse set of galaxies is capable of identifying major mergers, especially at early stages (74% accuracy), similar to that of networks trained at lower redshifts and/or higher masses (with accuracies between 66% and 80%). We discuss the inherent limits of galaxy merger identification due to the orientation angle, finding 98% of mergers are correctly identified from at least one angle, and 61% are correctly identified from the majority of angles. We additionally explore the confounding variables, such as star formation, to consider when applying to real data. This network enables the exploration of the impact of previously overlooked mergers of high-mass ratio and low stellar masses on galaxy evolution in CANDELS, and can be expanded to surveys from JWST, Rubin, Roman, and Euclid.