Non-parametric Lagrangian biasing from the insights of neural nets

TL;DR

This paper introduces a neural network-based Lagrangian bias model for galaxy clustering that uses multiple smoothing scales of the initial density field, improving the prediction of halo power spectra in simulations.

Contribution

It demonstrates how neural nets trained on multiple smoothing scales can effectively model galaxy bias, highlighting the importance of multi-scale features in bias modeling.

Findings

Including three smoothing scales yields the best halo power spectrum recovery.

Adding more scales can cause underestimation and overfitting.

Principal component analysis simplifies the feature space without losing accuracy.

Abstract

We present a Lagrangian model of galaxy clustering bias in which we train a neural net using the local properties of the smoothed initial density field to predict the late-time mass-weighted halo field. By fitting the mass-weighted halo field in the AbacusSummit simulations at z=0.5, we find that including three coarsely spaced smoothing scales gives the best recovery of the halo power spectrum. Adding more smoothing scales may lead to 2-5% underestimation of the large-scale power and can cause the neural net to overfit. We find that the fitted halo-to-mass ratio can be well described by two directions in the original high-dimension feature space. Projecting the original features into these two principal components and re-training the neural net either reproduces the original training result, or outperforms it with a better match of the halo power spectrum. The elements of the principal components are unlikely to be assigned physical meanings, partly owing to the features being highly correlated between different smoothing scales. Our work illustrates a potential need to include multiple smoothing scales when studying galaxy bias, and this can be done easily with machine-learning methods that can take in high dimensional input feature space.