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

TL;DR

This paper introduces a deep learning framework combining a volumetric CNN and an optimized Friends-of-Friends algorithm to identify dark matter haloes in cosmological simulations, achieving high accuracy and significantly faster processing than traditional methods.

Contribution

The authors develop a novel deep learning pipeline that improves the speed and scalability of dark matter halo identification in simulations, outperforming existing halo finders in accuracy and efficiency.

Findings

Achieved over 98% accuracy in classifying halo particles.

Halo catalogues with >95% purity and 93% completeness for high-mass haloes.

Speed-up of approximately ten times compared to ROCKSTAR.

Abstract

We present a deep-learning-based approach for identifying dark matter haloes in cosmological N-body simulations. Our framework consists of a volumetric Convolutional Neural Network to classify individual simulation particles as either halo or non-halo members, followed by a highly optimised and parallelised Friends-of-Friends clustering algorithm that groups the classified halo members into distinct haloes. The training data comprise simulations generated using GADGET-4, with labels obtained with the ROCKSTAR halo finder. Our models incorporate two main halo mass definitions, $M_{200\mathrm{b}}$ and $M_{\text{vir}}$, with similar performance. For haloes defined by the ROCKSTAR $M_{200\mathrm{b}}$ criterion, the classification network demonstrated stable performance across multiple simulation resolutions. For the highest resolution, it achieved over $98\%$ across all primary performance metrics when identifying halo particles. Furthermore, the FoF algorithm yielded halo catalogues with a purity generally exceeding $95\%$ and a stable completeness of $93\%$ for masses above $5\times10^{11} \, M_\odot$. Our pipeline recovered the centre-of-mass positions, velocities and halo masses with high fidelity, yielding a halo mass function consistent to within $5\%$ of the reference while faithfully reconstructing the internal density profiles. The primary objective of this study is to offer a faster and scalable alternative to conventional halo finders, achieving a speed-up of approximately one order of magnitude relative to ROCKSTAR, offering a promising pathway for modern simulation-based inference methods that rely on rapid and accurate structure identification.