ML in Astrophysical Turbulence I: Predicting Prestellar Cores in Magnetized Molecular Clouds using eXtreme Gradient Boosting

Nikhil P. S. Bisht, David C. Collins · 2026

Read ML in Astrophysical Turbulence I: Predicting Prestellar Cores in Magnetized Molecular Clouds using eXtreme Gradient Boosting by Nikhil P. S. Bisht, David C. Collins (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-08513.txt.

Introduction

Nikhil Bisht and David Collins train a gradient-boosted model to guess which gas parcels in stormy, magnetized clouds will form stars. Fed snapshots from cloud runs of place, speed, density, and field strength, it tracks parcels well and ranks collapse risk with high skill. Yet the point-wise view lacks the local context real collapse checks need, so the authors call for spatial designs.

Catalog record

CREATORNikhil P. S. Bisht, David C. Collins
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE40828
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.08513

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

ML in Astrophysical Turbulence I: Predicting Prestellar Cores in Magnetized Molecular Clouds using eXtreme Gradient Boosting Nikhil P. S. Bisht, David C. Collins Abstract Giant Molecular Clouds (GMCs) are dominated by supersonic turbulence, creating a complex network of shocks and filaments that regulate star formation. While the global inefficiency of star formation is well-observed, predicting exactly which gas parcels within a turbulent cloud will collapse to form stars remains a challenge. In this work, we present a supervised machine learning framework to forecast the Lagrangian history of prestellar cores in magnetohydrodynamic (MHD) turbulence. We utilize Extreme Gradient Boosting (XGBoost) to train a regression model on the trajectories of ∼ 2.1 \sim 2.1 million tracer particles evolved within a self-gravitating, turbulent MHD simulation. By mapping the instantaneous phase-space state (position, velocity, and density) of gas parcels to their future coordinates, our model successfully predicts the 3D evolution of star-forming cores over a horizon of ∼ 0.45 \sim 0.45 Myr ( 0.25 ​ t ff 0.25~t_{\rm ff} ). We achieve a global coefficient of determination of R 2 > 0.99 R^{2}>0.99 and demonstrate that the model captures the non-linear convergent flows characteristic of gravitational collapse. Crucially, we show that local phase-space information alone is sufficient to distinguish between transient density fluctuations and bound collapsing cores. This data-driven approach offers a computationally efficient alternative to traditional sink-particle algorithms and provides a pathway for developing high-fidelity subgrid models for galaxy-scale simulations. I Introduction The formation of stars is the fundamental driver of galaxy evolution, regulating the interstellar medium (ISM) through feedback and enriching the universe with heavy elements. Despite its importance, the precise mechanism that regulates the rate of star formation remains one of the central problems in astrophysics. Observations indicate that star formation is surprisingly inefficient; only 1\%-2\% of a molecular cloud’s mass is converted into stars per free-fall time ( \epsilon_{\rm ff}\approx 0.01 ) ( 18 ; 19 ) . This inefficiency is generally attributed to the interplay between self-gravity, which promotes collapse, and supersonic turbulence and magnetic fields, which provide support and shear apart potential star-forming regions ( 24 ; 15 ) . Giant Molecular Clouds (GMCs) are the primary nurseries for this process. Spanning scales from tens of parsecs ( \sim 10^{17} m) down to the size of prestellar cores at about 0.1 parsecs ( \sim 3\times 10^{15} m), these clouds represent a chaotic environment where gas spans orders of magnitude in density and temperature. The standard paradigm suggests that supersonic turbulence creates a complex network of shocks and filaments ( 20 ; 22 ) . Within this turbulent cascade, transient overdensities form and dissolve, but only those fluctuations that are gravitationally bound and magnetically supercritical (a rare subset) succeed in collapsing to form prestellar cores ( 31 ; 27 ; 11 ) . Once these cores collapse and protostars ignite, they inject significant energy back into their environment. Radiative feedback (e.g., UV heating and photoionization) and protostellar outflows play a critical role in regulating the global star formation efficiency by dispersing the surrounding gas and halting further accretion ( 24 ) . However, as this study focuses strictly on the kinematic origins and the initial turbulent fragmentation of the prestellar phase which is prior to the onset of stellar ignition, we approximate the gas thermodynamics using an isothermal equation of state and intentionally omit radiative feedback mechanisms. This isolates the purely gravity-driven turbulence and advective rules our machine learning model aims to learn. Figure 1: Evolution of the turbulent density field. We show the column density projection ( \Sigma ) of the simulation box at four evolutionary stages, ranging from the initial turbulent driving phase ( t=0 ) to the onset of widespread core collapse ( t\approx t_{\rm ff} ). Observational surveys using tracer molecules (e.g., {}^{12}\text{CO} , \text{NH}_{3} , \text{N}_{2}\text{H}^{+} ) and dust continuum emission have provided a wealth of data regarding the morphology of these regions ( 1 ; 13 ) . Recent high-resolution campaigns with ALMA and observational analysis of cloud statistics ( 6 ; 30 ) have highlighted the environmental dependence of core formation. However, observations suffer from projection effects (2D snapshots of 3D structures) and a lack of temporal evolution. We observe the ”before” (diffuse gas) and ”after” (protostars), but the critical dynamic phase of collapse is often obscured or ambiguous. To bridge this gap, high-resolution Magnetohydrodynamic (MHD) simulations have become indispensable. Codes such as Enzo , Flash , and Arepo allow us to model the non-linear equations of fluid dynamics coupled with gravity and magnetic fields ( 4 ; 16 ; 36 ) . While simulations provide full 6D phase-space information, identifying the Lagrangian history of star-forming gas, i.e., tracing the specific parcels of gas that will eventually become stars, remains a non-trivial data analysis challenge. Traditional methods often rely on thresholding (e.g., density cutoffs) or sink particles ( 2 ; 14 ; 35 ) , which identify collapse only after it has become irreversible. In recent years, Machine Learning (ML) has emerged as a powerful tool for analyzing the complex, high-dimensional datasets produced by astrophysical simulations. ML techniques have been successfully applied to estimate halo masses ( 26 ) , classify galaxy morphologies ( 12 ) , and infer physical parameters from turbulent fields ( 28 ) . In the context of star formation, recent works have begun to utilize Deep Learning to classify young stellar objects ( 34 ) and identify structures in molecular line data. However, most efforts focus on Convolutional Neural Networks (CNNs) applied to image data. In this work, we present a novel approach focusing on the Lagrangian dynamics of the gas. We treat the prediction of core collapse not as an image recognition task, but as a tabular supervised learning problem. We utilize XGBoost (Extreme Gradient Boosting), a decision-tree-based ensemble method ( 5 ) , to analyze the trajectories of passive tracer particles in a turbulent MHD simulation. Unlike ”black box” neural networks, tree-based models are highly effective for tabular physical data (density, velocity, magnetic field) and offer interpretability regarding feature importance. Our goal is to answer a specific question: Given the instantaneous state of a gas parcel in a turbulent cloud, can we predict its future coordinate and accretion fate? This paper is structured as follows. Section II details the MHD simulation setup, the tracer particle methodology, and the data pipeline. Section III describes the problem formulation and the XGBoost architecture. Section IV presents the model performance, including temporal generalization tests and core trajectory reconstruction. We discuss the physical implications of our findings in Section V and conclude in Section VI . Table 1: Dataset MCDS1 ID Frame x_{i} y_{i} z_{i} v_{x,i} v_{y,i} v_{z,i} \rho_{i} x_{f} y_{f} z_{f} [code units] [code units] [code units] [code units] 4 20 0.957 0.049 0.020 -9.77 5.67 2.12 0.17 0.860 0.133 0.041 5 20 0.969 0.049 0.021 -9.18 5.61 2.21 0.57 0.871 0.124 0.050 6 20 0.982 0.049 0.021 -8.68 5.69 2.41 0.75 0.884 0.118 0.057 15 20 0.055 0.044 0.038 -8.73 4.82 4.90 0.45 0.966 0.087 0.099 17 20 0.079 0.035 0.043 -7.94 3.90 5.31 0.46 0.991 0.081 0.112 II Simulations and Data II.1 Turbulence Box Simulations We utilize the open-source adaptive mesh refinement (AMR) code Enzo ( 4

Read the complete text