{"ID":2888495,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.00049","arxiv_id":"2508.00049","title":"Segmenting proto-halos with vision transformers","abstract":"The formation of dark-matter halos from small cosmological perturbations generated in the early universe is a highly non-linear process typically modeled through N-body simulations. In this work, we explore the use of deep learning to segment and classify proto-halo regions in the initial density field according to their final halo mass at redshift z=0. We compare two architectures: a fully convolutional neural network (CNN) based on the V-Net design and a U-Net transformer. We find that the transformer-based network significantly outperforms the CNN across all metrics, achieving sub-percent error in the total segmented mass per halo class. Both networks deliver much higher accuracy than the perturbation-theory-based model \\textsc{pinocchio}, especially at low halo masses and in the detailed reconstruction of proto-halo boundaries. We also investigate the impact of different input features by training models on the density field, the tidal shear, and their combination. Finally, we use Grad-CAM to generate class-activation heatmaps for the CNN, providing preliminary yet suggestive insights into how the network exploits the input fields.","short_abstract":"The formation of dark-matter halos from small cosmological perturbations generated in the early universe is a highly non-linear process typically modeled through N-body simulations. In this work, we explore the use of deep learning to segment and classify proto-halo regions in the initial density field according to the...","url_abs":"https://arxiv.org/abs/2508.00049","url_pdf":"https://arxiv.org/pdf/2508.00049v2","authors":"[\"Toka Alokda\",\"Cristiano Porciani\"]","published":"2025-07-31T17:59:44Z","proceeding":"astro-ph.CO","tasks":"[\"astro-ph.CO\",\"astro-ph.IM\",\"cs.CV\"]","methods":"[\"Vision Transformer\",\"Transformer\",\"Convolutional Neural Network\"]","has_code":false}