Woo, NathanWang, ZhuoyiMayotte, EricMayotte, Sonja2024-05-082024-05-08https://hdl.handle.net/11124/179064https://doi.org/10.25676/11124/179064Ultra-high-energy cosmic rays are single atomic nuclei from other galaxies and the most energetic phenomena known to humankind. In observing these and using them to increase our astrophysical understanding, two properties are of key importance: the particle's energy and its mass. When one of these particles strikes the top of the atmosphere, they are destroyed and create a high-energy particle cascade which can reach 10 billion particles in size. Here, we present a new machine learning method to maximize the amount of information that can be extracted on the mass of the ultra-high-energy cosmic ray. We find that a neural networks such as a convolutional neural network can improve the mass prediction of ultra-high-energy cosmic rays, with a proton-iron merit factor of 9.81, three times better than traditional solutions. The extra precision on mass reconstruction for these cosmic rays can lead to a revolution on the quality of astrophysics performed with current ground-based observatories, and displays the usefulness of machine learning techniques in astrophysics.postersengCopyright of the original work is retained by the author.Text