SciPost Submission Page
How to GAN away Detector Effects
by Marco Bellagente, Anja Butter, Gregor Kasieczka, Tilman Plehn, Ramon Winterhalder
- Published as SciPost Phys. 8, 070 (2020)
|As Contributors:||Tilman Plehn · Ramon Winterhalder|
|Arxiv Link:||https://arxiv.org/abs/1912.00477v4 (pdf)|
|Date submitted:||2020-03-19 01:00|
|Submitted by:||Winterhalder, Ramon|
|Submitted to:||SciPost Physics|
|Subject area:||High-Energy Physics - Phenomenology|
LHC analyses directly comparing data and simulated events bear the danger of using first-principle predictions only as a black-box part of event simulation. We show how simulations, for instance, of detector effects can instead be inverted using generative networks. This allows us to reconstruct parton level information from measured events. Our results illustrate how, in general, fully conditional generative networks can statistically invert Monte Carlo simulations. As a technical by-product we show how a maximum mean discrepancy loss can be staggered or cooled.
Ontology / TopicsSee full Ontology or Topics database.
Published as SciPost Phys. 8, 070 (2020)
Submission & Refereeing History
You are currently on this page
Reports on this Submission
Anonymous Report 2 on 2020-4-11 Invited Report
- Cite as: Anonymous, Report on arXiv:1912.00477v4, delivered 2020-04-11, doi: 10.21468/SciPost.Report.1620
I am satisfied with authors replies to my previous comments.
The paper overall has improved (also thanks to other referees comments) and I think it is already in good shape to be published.
I appreciated in particular the new section 4 "New physics injection": it is not mandatory, but if there will be a new submission, I think it would be nice to see also the detector distributions in Figure 10 (in particular I am interested to see it for m_lljj, where I wonder if the unfolded new resonance, without additional constraints, improves its significance over the detector level one).
The hyperlink is missing in Ref 23: G. Cowan, Conf. Proc. C0203181 (2002) 248. [,248(2002)]
(also "[,248(2002)]" should not appear)
Many congratulations to the authors for this very nice and inspiring work.
Anonymous Report 1 on 2020-3-29 Invited Report
- Cite as: Anonymous, Report on arXiv:1912.00477v4, delivered 2020-03-29, doi: 10.21468/SciPost.Report.1599
I'd like to thank the authors for their responses and the clarifications in the revised version. I appreciate the extra test of the robustness of the method in the presence of signatures that are significantly different from the Standard Model. I think that makes for a nice addition to the paper.