SciPost Submission Page
Generative Unfolding with Distribution Mapping
by Anja Butter, Sascha Diefenbacher, Nathan Huetsch, Vinicius Mikuni, Benjamin Nachman, Sofia Palacios Schweitzer, Tilman Plehn
This Submission thread is now published as
Submission summary
Authors (as registered SciPost users): | Nathan Huetsch · Sofia Palacios Schweitzer · Tilman Plehn |
Submission information | |
---|---|
Preprint Link: | scipost_202504_00043v1 (pdf) |
Date accepted: | May 26, 2025 |
Date submitted: | April 28, 2025, 3:35 p.m. |
Submitted by: | Huetsch, Nathan |
Submitted to: | SciPost Physics |
Ontological classification | |
---|---|
Academic field: | Physics |
Specialties: |
|
Approaches: | Experimental, Computational, Phenomenological |
Abstract
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, Schr\"odinger Bridges and Direct Diffusion, in order to ensure that the models learn the correct conditional probabilities. This brings distribution mapping 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.
Author indications on fulfilling journal expectations
- Provide a novel and synergetic link between different research areas.
- Open a new pathway in an existing or a new research direction, with clear potential for multi-pronged follow-up work
- Detail a groundbreaking theoretical/experimental/computational discovery
- Present a breakthrough on a previously-identified and long-standing research stumbling block
Author comments upon resubmission
Dear editor, dear reviewers,
Thank you very much for taking the time to review our work and for your positive feedback.
We carefully went through the list of suggested changes and incorporated them. We fixed the typos and slightly reformulated the statements that were misleading. For a point-by-point breakdown, see the list of changes.
Thank you once again for reviewing our publication.
On behalf of the authors
Nathan Huetsch
Thank you very much for taking the time to review our work and for your positive feedback.
We carefully went through the list of suggested changes and incorporated them. We fixed the typos and slightly reformulated the statements that were misleading. For a point-by-point breakdown, see the list of changes.
Thank you once again for reviewing our publication.
On behalf of the authors
Nathan Huetsch
List of changes
Fixed some typos. In addition we made the following changes:
1)
p2, 2nd paragraph: Reformulated to "These results are promising, however due to the complexity of the problem, it is essential to have orthogonal methods that can cross-check each other."
2)
p2, last paragraph: Deleted the misleading statement " ... as well as any unfolding method."
3)
p4, first paragraph: Here we were referring to the dataset published alongside the first Omnifold paper [arXiv:1911.09107], because it is often used as a benchmark for ML-based unfolding methods. We were not referring to Omnifold as a method. Nevertheless, we changed "standard" to "common".
4)
We increased the label size in Fig.8
1)
p2, 2nd paragraph: Reformulated to "These results are promising, however due to the complexity of the problem, it is essential to have orthogonal methods that can cross-check each other."
2)
p2, last paragraph: Deleted the misleading statement " ... as well as any unfolding method."
3)
p4, first paragraph: Here we were referring to the dataset published alongside the first Omnifold paper [arXiv:1911.09107], because it is often used as a benchmark for ML-based unfolding methods. We were not referring to Omnifold as a method. Nevertheless, we changed "standard" to "common".
4)
We increased the label size in Fig.8
Published as SciPost Phys. 18, 200 (2025)