Towards a data-driven model of hadronization using normalizing flows

Abstract

We introduce a model of hadronization based on invertible neural networks that faithfully reproduces a simplified version of the Lund string model for meson hadronization. Additionally, we introduce a new training method for normalizing flows, termed MAGIC, that improves the agreement between simulated and experimental distributions of high-level (macroscopic) observables by adjusting single-emission (microscopic) dynamics. Our results constitute an important step toward realizing a machine-learning based model of hadronization that utilizes experimental data during training. Finally, we demonstrate how a Bayesian extension to this normalizing-flow architecture can be used to provide analysis of statistical and modeling uncertainties on the generated observable distributions.

arXiv comment26 pages, 9 figures, public code available

References in corpus (16)

• Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC
• Herwig++ Physics and Manual
• Normalizing Flows: An Introduction and Review of Current Methods
• Herwig 7.0 / Herwig++ 3.0 Release Note
• Tuning PYTHIA 8.1: the Monash 2013 Tune
• Bayesian Workflow
• Neural Networks for Full Phase-space Reweighting and Parameter Tuning
• Thermodynamical String Fragmentation
• Generative Networks for Precision Enthusiasts
• The Hidden Geometry of Particle Collisions
• Event generator tuning using Bayesian optimization
• Conditional Density Estimation with Bayesian Normalising Flows
• Towards a Deep Learning Model for Hadronization
• Apprentice for Event Generator Tuning
• Fragmentation Uncertainties in Hadronic Observables for Top-quark Mass Measurements
• Reweighting Monte Carlo Predictions and Automated Fragmentation Variations in Pythia 8

Cited by in corpus (10)

• CaloDREAM -- Detector Response Emulation via Attentive flow Matching
• A Lorentz-Equivariant Transformer for All of the LHC
• Normalizing Flows for High-Dimensional Detector Simulations
• Convolutional L2LFlows: Generating Accurate Showers in Highly Granular Calorimeters Using Convolutional Normalizing Flows
• CaloChallenge 2022: A Community Challenge for Fast Calorimeter Simulation
• Moment Unfolding
• Describing Hadronization via Histories and Observables for Monte-Carlo Event Reweighting
• Unifying Simulation and Inference with Normalizing Flows
• Data-Driven Predictions for Dark Photon and Millicharged Particle Production
• Characterizing the hadronization of parton showers using the HOMER method