Cover Image

Generative unfolding with distribution mapping

Machine learning enables unbinned, highly-differential cross section measurements. A recent idea uses generative models to morph a starting simulation into the unfolded data. We show how to extend two morphing techniques, Schrödinger Bridges and Direct Diffusion, in order to ensure that the models...

Full description

Saved in:
Bibliographic Details
Main Authors: Butter, Anja (Author) , Diefenbacher, Sascha (Author) , Hütsch, Nathan (Author) , Mikuni, Vinicius (Author) , Nachman, Benjamin (Author) , Palacios Schweitzer, Sofia (Author) , Plehn, Tilman (Author)
Format: Article (Journal)
Language:English
Published: June 2025
In: SciPost physics
Year: 2025, Volume: 18, Issue: 6, Pages: 1-28
ISSN:2542-4653
DOI:10.21468/SciPostPhys.18.6.200
Online Access:Verlag, kostenfrei, Volltext: https://doi.org/10.21468/SciPostPhys.18.6.200
Verlag, kostenfrei, Volltext: https://scipost.org/10.21468/SciPostPhys.18.6.200
Get full text
Author Notes:Anja Butter, Sascha Diefenbacher, Nathan Huetsch, Vinicius Mikuni, Benjamin Nachman, Sofia Palacios Schweitzer and Tilman Plehn
Description
Summary:Machine learning enables unbinned, highly-differential cross section measurements. A recent idea uses generative models to morph a starting simulation into the unfolded data. We show how to extend two morphing techniques, Schrödinger Bridges and Direct Diffusion, in order to ensure that the models learn the correct conditional probabilities. This brings distribution mapping (DM) to a similar level of accuracy as the state-of-the-art conditional generative unfolding methods. Numerical results are presented with a standard benchmark dataset of single jet substructure as well as for a new dataset describing a 22-dimensional phase space of Z+2-jets.
Item Description:Online erschienen: 20. Juni 2025
Gesehen am 15.12.2025
Physical Description:Online Resource
ISSN:2542-4653
DOI:10.21468/SciPostPhys.18.6.200

Similar Items