Normalizing Flows to Reconstruct Pseudo-PDFs

2026/07/28 by Yamil Cahuana Medrano, Kostas Orginos
#hep-lat #cs.LG #hep-ph

paper · pdf

Abstract

We investigate a normalizing-flow approach for reconstructing parton distribution functions (PDFs) from synthetic matrix-element data. Our framework combines Gaussian Process priors with invertible neural networks to learn a posterior distribution over PDFs consistent with limited Ioffe-time data. We demonstrate that the architecture preserves physical constraints and extrapolation properties.

Citations

• Gaussian Processes for Inferring Parton Distributions
• Quantization Of Probability Measures In Maximum~Mean~Discrepancy Distance
• Bayesian Inference with Gaussian Processes for the Determination of Parton Distribution Functions
• Analyzing Inverse Problems with Invertible Neural Networks
• SGDR: Stochastic Gradient Descent with Warm Restarts
• Density estimation using Real NVP

Related

• Parton distribution functions
• A linear PDF model for Bayesian inference
• Normalizing Flows for Domain Adaptation when Identifying Λ Hyperon Events
• Bayesian Inference with Gaussian Processes for the Determination of Parton Distribution Functions
• An open-source machine learning framework for global analyses of parton distributions