CNN+FoF: application of deep learning to the identification of dark matter haloes

Name: CNN+FoF: application of deep learning to the identification of dark matter haloes
Uploaded: 2026-10-03T21:29:50.196203+00:00
Duration: 536 s
Description: This framework uses a deep learning model (CNN) to classify simulation particles as halo or non-halo members, followed by an optimized Friends-of-Friends (FoF) clustering algorithm. This hybrid method provides a one-order-of-magnitude speedup over ROCKSTAR while maintaining high accuracy in identifying dark matter haloes and recovering their spatial and kinematic properties.

This framework uses a deep learning model (CNN) to classify simulation particles as halo or non-halo members, followed by an optimized Friends-of-Friends (FoF) clustering algorithm. This hybrid method provides a one-order-of-magnitude speedup over ROCKSTAR while maintaining high accuracy in identifying dark matter haloes and recovering their spatial and kinematic properties.

Open the full episode