A new method for the q2 reconstruction in semileptonic decays at LHCb based on machine learning - Abstract

Aug 5, 2026

Upload by Jiven

The neutrino closure method is commonly used to derive semileptonic decay kinematics when one particle remains unreconstructed, but the resulting quadratic relation introduces a two-fold ambiguity. A machine-learning approach resolves this ambiguity by selecting suitable feature sets and regression models to improve the reconstructed invariant mass squared q2 of the lν system. The study shows the best resolution using support vector features with a multilayer perceptron regressor, yielding about a 40% improvement versus random selection, and demonstrates applicability across various semileptonic decays.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages13

Reading Time33 min

AccessFree to download

Abstract

Introduction

arXiv :2208 .02145v1 [hep-ph] 3 Aug 2022 A new method for the q 2 reconstruction in semileptonic decays at LHCb based on machine learning Panting Ge, 1 Xiaotao Huang,2 Miroslav Saur,3 and Liang Sun 1, 􀀃 1 School of Physics and Technology, Wuhan University, Wuhan 430072, China 2 The Institute for Advanced Studies, Wuhan University, Wuhan 430072, China 3 Technische Universit􀁿at Dortmund, Dortmund 44227, Germany (Dated: Received: date / Accepted: date) Abstract The neutrino closure method is often used to obtain kinematics of semileptonic decays with one unreconstructed particle. The kinematics of decays can be deducted by a two-fold ambiguity with a quadratic equation. To resolve the two-fold ambiguity, a new method based on Machine Learning (ML) is proposed. We study the e􀀋ect of di􀀋erent sets of features and regressors on the improvement of reconstructed invariant mass squared of `􀀗 system (q2 ) . The result shows that the best performance is obtained by using the 􀀍ight vector as the features, and the multilayer perceptron (MLP) model as the regressor. Compared with the random choice, the MLP model improves the resolution of reconstructed q 2 by 􀀘 40% . Furthermore, the possibility of using this method on various semileptonic decays is shown. 􀀃 Corresponding author: [sunl@whu.edu.cn](sunl@whu.edu.cn) I. INTRODUCTION Semileptonic decays, mediated by a virtual W boson which produces one lepton and the corresponding neutrino in addition to one or more hadrons, o􀀋er a good platform to study the weak as well as strong interaction e􀀋ects [1] . Studies of semileptonic decays, therefore, have been paid much more attention in recent years, especially for the purposes of precise measurements on the Cabibbo-Kobayashi-Maskawa (CKM) matrix elements, such as the determination of j Vubj and jVcbj . The precision measurement of the CKM matrix elements allows us: I) to test the unitarity of the CKM matrix [2] and then to measure the angle 􀀌; II) to improve the precision on the least well-known CKM elements; III) to help resolve the tension between exclusive and inclusive measurements; IV) to probe new physics (NP); and V) to help predict other branching fractions, such as B ! 􀀜􀀗􀀜 . Additionally, the recent measurements of the branching fraction ratios R(D? ) = B(B ! D(?)􀀜􀀗)=B(B ! D (?)􀀖􀀗) measured in experiments show a slight disagreement with the Standard Model (SM) predictions [3] . Based on the above, the studies of semileptonic decays by LHCb experiment, which focuses on a heavy-􀀍avour studies in a forward region, show an increasing trend, although the presence of an unreconstructed neutrino is experimentally challenging. At B-factories operating at the 􀀇(4S) resonance, the kinematics of missing particles 􀀖 in B mesons can be reconstructed by balancing against the B decay [4], while in hadron collider experiments the studies of semileptonic decays pose a technical challenge due to the unreconstructed neutrino in the 􀀌nal state. Hadron collider experiments provide a great number of opportunities to precisely determine CKM matrix elements in semileptonic decays, although some challenges still remain in resolving the missing neutrino due to the busy hadronic environment and the production mechanism. First of all, a large Lorentz boost can be produced by hadron collider experiments, especially at the forward rapidity covered by the LHCb experiment [5], which is one of the major experiments at LHC. Secondly, the decay kinematics can be restricted by the b-hadron decay vertex and the measured 􀀍ight vector which connects with the primary pp interaction vertex [6] . Finally, the mass of single missing particles can be deduced from the conservation of four-momentum. Conservation of the transverse momentum to the 􀀍ight vector provides two independent constraints on the semileptonic decays as well. A third constraint is the parent b-hadron mass should be conserved, though this condition has an ambiguity which produ

FAQs

Why is q2 reconstruction in semileptonic decays difficult at hadron colliders like LHCb?

Because the neutrino is unreconstructed in the final state, so its momentum must be inferred using kinematic constraints, which creates an intrinsic two-solution ambiguity.

What does the proposed method use to resolve the two-fold ambiguity?

It replaces manual resolution with a machine-learning regression strategy, testing different feature sets and regressors to improve the reconstructed q2 resolution.

Which machine-learning configuration gives the best performance, and what improvement is reported?

Using support vector features with a multilayer perceptron regressor provides the best results, improving the q2 resolution by about 40% compared with random choice.

Aug 5, 2026

Upload by Jiven