Potential and limitations of machine-learning approaches to inclusive determinations
arXiv:2105.09271 · doi:10.1007/JHEP01(2022)143
Abstract
The determination of in inclusive semileptonic decays will be among the pivotal tasks of Belle II. In this paper we study the potential and limitations of machine learning approaches that attempt to reduce theory uncertainties by extending the experimentally accessible fiducial region of the signal into regions where the background is dominant. We find that a deep neural network trained on low-level single particle features offers modest improvement in separating signal from background, compared to BDT set-ups using physicist-engineered high-level features. We further illustrate that while the signal acceptance of such a deep neural network deteriorates in kinematic regions where the signal is small, such as at high hadronic invariant mass, neural networks which exclude kinematic features are flatter in kinematics but less inclusive in the sampling of exclusive hadronic final states at fixed kinematics. The trade-off between these two set-ups is somewhat Monte Carlo dependent, and we study this issue using the multipurpose event generator Sherpa in addition to the widely used -physics tool EvtGen.
arXiv comment20 pages, 11 figures; v2: Final version published in JHEP
References in corpus (41)
- Adam: A Method for Stochastic Optimization
- XGBoost: A Scalable Tree Boosting System
- An Introduction to PYTHIA 8.2
- A Brief Introduction to PYTHIA 8.1
- Event generation with SHERPA 1.1
- Searching for Exotic Particles in High-Energy Physics with Deep Learning
- Tuning PYTHIA 8.1: the Monash 2013 Tune
- Inclusive Semileptonic B and Polarized Lambda_b Decays from QCD
- Nonperturbative Corrections to Inclusive Beauty and Charm Decays: QCD versus Phenomenological Models
- Deep Learning and its Application to LHC Physics
- Differential Distributions in Semileptonic Decays of the Heavy Flavors in QCD
- On the Motion of Heavy Quarks inside Hadrons: Universal Distributions and Inclusive Decays
- TensorFlow Distributions
- Factorization and Shape-Function Effects in Inclusive B-Meson Decays
- QCD-Based Interpretation of the Lepton Spectrum in Inclusive Decays
- Theory of Charmless Inclusive B Decays and the Extraction of V_{ub}
- Jet Flavor Classification in High-Energy Physics with Deep Neural Networks
- Flipout: Efficient Pseudo-Independent Weight Perturbations on Mini-Batches
- Determination of the quark coupling strength using baryonic decays
- Study of B --> pi l nu and B --> rho l nu Decays and Determination of |V_ub|
- Study of Exclusive B->X_u l nu Decays and Extraction of |V_{ub}| using Full Reconstruction Tagging at the Belle Experiment
- B -> X_u l nu decay distributions to order alpha_s
- Inclusive semileptonic B decays and the determination of |V_ub|
- Inclusive spectra in charmless semileptonic B decays by Dressed Gluon Exponentiation
- Measurement of the decay and determination of
- Branching fraction and form-factor shape measurements of exclusive charmless semileptonic B decays, and determination of |V_{ub}|
- Can the differences in the determinations of and be explained by New Physics?
- Multivariate discrimination and the Higgs + W/Z search
- corrections to fully-differential top quark decays
- Subleading Shape Functions in Inclusive B Decays
- The second order QCD contribution to the semileptonic b -> u decay rate
- Factorization for Power Corrections to B->Xs gamma and B-> Xu e nu
- Power corrections to B\to X_u l ν(X_sγ) decay spectra in the "shape-function" region
- Deep-Learning Jets with Uncertainties and More
- Measurement Of |V_ub| From Inclusive Charmless Semileptonic B Decays
- Study of Bbar --> Xu l nubar decays in BBbar events tagged by a fully reconstructed B-meson decay and determination of |V_{ub}|
- On the corrections to inclusive decays
- NNLO corrections to B --> X_u l nu and the determination of |V_ub|
- Perturbative corrections to power suppressed effects in
- The first year of the BABAR experiment at PEP-II
- Vub extraction using the Analytic Coupling model