Machine learning classification: Case of Higgs boson CP state in H→ττ decay at the LHC

Abstract

Machine Learning (ML) techniques are rapidly finding a place among the methods of High Energy Physics data analysis. Different approaches are explored concerning how much effort should be put into building highlevel variables based on physics insight into the problem, and when it is enough to rely on low-level ones, allowing ML methods to find patterns without explicit physics model.In this paper we continue the discussion of previous publications on the CP state of the Higgs boson measurement of the H → ττ decay channel with the consecutive τ ± → ρ ± ν; ρ ± → π ± π 0 and τ ± → a ± 1 ν; a ± 1 → ρ 0 π ± → 3π ± cascade decays. The discrimination of the Higgs boson CP state is studied as a binary classification problem between CP-even (scalar) and CP-odd (pseudoscalar), using Deep Neural Network (DNN). Improvements on the classification from the constraints on directly non-measurable outgoing neutrinos are discussed. We find, that once added, they enhance the sensitivity sizably, even if only imperfect information is provided. In addition to DNN we also evaluate and compare other ML methods: Boosted Trees (BT), Random Forest (RF) and Support Vector Machine (SVN).

Extract the energy scale of anomalous γγ → W+W− scattering in the vector boson scattering process using artificial neural networks

“…In response to the problems in determining the polarizations, a novel approach has been introduced into high energy physics (HEP). It has been shown that, the artificial neural network (ANN) can be very powerful in determining the polarizations of W, Z bosons [21][22][23] and τ lepton [24]. The ANN approach is one of the machine learning methods, which have been widely used in HEP, and are being developed rapidly in recent years [25][26][27][28][29][30][31][32][33][34].…”

Application of artificial intelligence in the determination of impact parameter in heavy-ion collisions at intermediate energies

“…Recently, the field of artificial intelligence (AI) has received unprecedented attention, and prodigious progress has been made in the application of the AI techniques [33][34][35][36]. Particularly, the neural network framework that is based on the simulation of the brain has been widely used in many physical researches [37][38][39][40][41][42][43][44][45][46][47][48][49][50][51]. Thus, it is timely to deduce the impact parameter with the new developed machine learning and deep learning algorithms.…”

Deep Neural Network application: Higgs boson CP state mixing angle in H to tau tau decay and at LHC

“…It can be implemented in the Monte Carlo simulations as per event spin weight wt, see [15] for more details. In [16,17] we have performed analysis for the three channels of the τ lepton-pair decays, respectively ρ ± ν τ ρ ∓ ν τ , a ± 1 ν τ ρ ∓ ν τ and a ± 1 ν τ a ∓ 1 ν τ but we limited the scope to the scalar-pseudoscalar classification case. We explored the kinematics of outgoing decay products of the τ leptons and geometry of decay vertices.…”

“…We analyze possible solutions with Deep Neural Network (DNN) algorithms [4] implemented in Tensorflow environment [18] which we have previously found working well for the binary classification [16,17] between scalar or pseudoscalar Higgs boson variants (which correspond to φ CP = 0 and φ CP = π/2). Our goals for the DNN algorithms is to determine per event:…”

“…In Section 3 we review lists of features (variables) used as an input to DNN and present samples prepared for analyses. As a straightforward extension of [16,17], still using binary classification, we analyze possibility to distinguish between scalar and arbitrary mix of scalar/pseudoscalar states. This study is covered in Section 4.…”