# A volumetric deep Convolutional Neural Network for simulation of mock dark matter halo catalogues

arXiv: https://arxiv.org/abs/1805.04537 · HTML: https://mlsanity.com/paper/1805.04537/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 11 May 2018.

## Built on by (most cited first, 7 of 7)

- [Dimensionless machine learning: Imposing exact units equivariance](https://mlsanity.com/paper/2204.00887/) (2022)
- [Astronomia ex machina: a history, primer, and outlook on neural networks in astronomy](https://mlsanity.com/paper/2211.03796/) (2022)
- [Field Level Neural Network Emulator for Cosmological N-body Simulations](https://mlsanity.com/paper/2206.04594/) (2022)
- [Augmenting astrophysical scaling relations with machine learning: application to reducing the Sunyaev-Zeldovich flux-mass scatter](https://mlsanity.com/paper/2201.01305/) (2022)
- [AI-assisted super-resolution cosmological simulations II: Halo substructures, velocities and higher order statistics](https://mlsanity.com/paper/2105.01016/) (2021)
- [Segmenting proto-halos with vision transformers](https://mlsanity.com/paper/2508.00049/) (2025)
- [Emulating Cosmic Structure Formation with a Lagrangian Neural Cellular Automaton](https://mlsanity.com/paper/2607.27320/) (2026)

## Abstract (arXiv)

For modern large-scale structure survey techniques it has become standard practice to test data analysis pipelines on large suites of mock simulations, a task which is currently prohibitively expensive for full N-body simulations. Instead of calculating this costly gravitational evolution, we have trained a three-dimensional deep Convolutional Neural Network (CNN) to identify dark matter protohalos directly from the cosmological initial conditions. Training on halo catalogues from the Peak Patch semi-analytic code, we test various CNN architectures and find they generically achieve a Dice coefficient of ~92% in only 24 hours of training. We present a simple and fast geometric halo finding algorithm to extract halos from this powerful pixel-wise binary classifier and find that the predicted catalogues match the mass function and power spectra of the ground truth simulations to within ~10%. We investigate the effect of long-range tidal forces on an object-by-object basis and find that the network's predictions are consistent with the non-linear ellipsoidal collapse equations used explicitly by the Peak Patch algorithm.