Search papers, labs, and topics across Lattice.
This paper explores a novel normalizing-flow method to reconstruct parton distribution functions (PDFs) from synthetic matrix-element data, leveraging Gaussian Process priors alongside invertible neural networks. The approach is significant as it maintains physical constraints and extrapolation properties while learning a posterior distribution over PDFs, even with limited Ioffe-time data. Key results indicate that the proposed framework effectively captures the underlying physics, enhancing the accuracy of PDF reconstruction in high-energy physics contexts.
Normalizing flows can accurately reconstruct parton distribution functions while respecting physical constraints, even from sparse data.
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.
