Inferring identified hadron production in $pp$ collisions with physics-informed machine learning at the LHC

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Gupta, Rishabh, Goswami, Kangkan, Prasad, Suraj, Sahoo, Raghunath
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866918492658204672
author Gupta, Rishabh
Goswami, Kangkan
Prasad, Suraj
Sahoo, Raghunath
author_facet Gupta, Rishabh
Goswami, Kangkan
Prasad, Suraj
Sahoo, Raghunath
contents Machine learning has become a powerful tool in high-energy collider experiments, which enables the studies based on data-driven approaches to complex reconstruction and regression tasks. The study of identified hadron spectra in pseudorapidity regions beyond detector acceptance, which is limited to mid-rapidity regions, carries important information about particle production, yet remains unmeasured. In this work, we develop a physics-informed neural network, trained on PYTHIA8 $pp$ collisions at $\sqrt{s}=13.6$ TeV, to infer $p_{\rm T}$ spectra of $π^{\pm}$, $K^{\pm}$, $p/\bar{p}$, $Λ/\barΛ$, and $K^{0}_{\mathrm{s}}$ in different rapidity regions. Physics-motivated constraints, including particle yield ratios, spectral shape, and smoothness, are incorporated into the loss function. A staged hyperparameter optimization strategy is used to ensure stability. The model achieves yield uncertainties of ${\sim}1.5\%$, $1.8\%$, and $5.83\%$ in the training, interpolation, and extrapolation regimes, respectively, outperforming XGBoost and LightGBM. It further reproduces key observables such as particle yield ratios, the multiplicity dependence of $\langle p_{\rm T} \rangle$, and kinetic freeze-out parameters, indicating that the model captures the underlying physics and provides reliable predictions beyond the measured phase space.
format Preprint
id arxiv_https___arxiv_org_abs_2605_09022
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Inferring identified hadron production in $pp$ collisions with physics-informed machine learning at the LHC
Gupta, Rishabh
Goswami, Kangkan
Prasad, Suraj
Sahoo, Raghunath
High Energy Physics - Phenomenology
High Energy Physics - Experiment
High Energy Physics - Theory
Nuclear Experiment
Nuclear Theory
Machine learning has become a powerful tool in high-energy collider experiments, which enables the studies based on data-driven approaches to complex reconstruction and regression tasks. The study of identified hadron spectra in pseudorapidity regions beyond detector acceptance, which is limited to mid-rapidity regions, carries important information about particle production, yet remains unmeasured. In this work, we develop a physics-informed neural network, trained on PYTHIA8 $pp$ collisions at $\sqrt{s}=13.6$ TeV, to infer $p_{\rm T}$ spectra of $π^{\pm}$, $K^{\pm}$, $p/\bar{p}$, $Λ/\barΛ$, and $K^{0}_{\mathrm{s}}$ in different rapidity regions. Physics-motivated constraints, including particle yield ratios, spectral shape, and smoothness, are incorporated into the loss function. A staged hyperparameter optimization strategy is used to ensure stability. The model achieves yield uncertainties of ${\sim}1.5\%$, $1.8\%$, and $5.83\%$ in the training, interpolation, and extrapolation regimes, respectively, outperforming XGBoost and LightGBM. It further reproduces key observables such as particle yield ratios, the multiplicity dependence of $\langle p_{\rm T} \rangle$, and kinetic freeze-out parameters, indicating that the model captures the underlying physics and provides reliable predictions beyond the measured phase space.
title Inferring identified hadron production in $pp$ collisions with physics-informed machine learning at the LHC
topic High Energy Physics - Phenomenology
High Energy Physics - Experiment
High Energy Physics - Theory
Nuclear Experiment
Nuclear Theory
url https://arxiv.org/abs/2605.09022