Jet Diffusion versus JetGPT -- Modern Networks for the LHC

Abstract

We introduce two diffusion models and an autoregressive transformer for LHC physics simulations. Bayesian versions allow us to control the networks and capture training uncertainties. After illustrating their different density estimation methods for simple toy models, we discuss their advantages for Z plus jets event generation. While diffusion networks excel through their precision, the transformer scales best with the phase space dimensionality. Given the different training and evaluation speed, we expect LHC physics to benefit from dedicated use cases for normalizing flows, diffusion models, and autoregressive transformers.

37 pages, 18 figures. v2: added appendix A, fixed typos

Cited by in corpus (15)

• How to Understand Limitations of Generative Networks
• The MadNIS Reloaded
• Masked Particle Modeling on Sets: Towards Self-Supervised High Energy Physics Foundation Models
• Deep Generative Models for Detector Signature Simulation: A Taxonomic Review
• High-dimensional and Permutation Invariant Anomaly Detection
• The Landscape of Unfolding with Machine Learning
• Precision-Machine Learning for the Matrix Element Method
• Normalizing Flows for High-Dimensional Detector Simulations
• Convolutional L2LFlows: Generating Accurate Showers in Highly Granular Calorimeters Using Convolutional Normalizing Flows
• Differentiable MadNIS-Lite
• Calibrating Bayesian Generative Machine Learning for Bayesiamplification
• Learning the Simplicity of Scattering Amplitudes
• Generative Unfolding with Distribution Mapping
• How to Unfold Top Decays
• TrackFormers: In Search of Transformer-Based Particle Tracking for the High-Luminosity LHC Era