SciPost Submission Page
Automatizing the search for mass resonances using BumpNet
by Jean-François Arguin, Georges Azuelos, Émile Baril, Ilan Bessudo, Fannie Bilodeau, Maryna Borysova, Shikma Bressler, Samuel Calvet, Julien Donini, Etienne Dreyer, Michael Kwok Lam Chu, Eva Mayer, Ethan Meszaros, Nilotpal Kakati, Bruna Pascual Dias, Joséphine Potdevin, Amit Shkuri, Muhammad Usman
Submission summary
| Authors (as registered SciPost users): | Ethan Meszaros |
| Submission information | |
|---|---|
| Preprint Link: | https://arxiv.org/abs/2509.16282v1 (pdf) |
| Date submitted: | Sept. 24, 2025, 4:50 a.m. |
| Submitted by: | Ethan Meszaros |
| Submitted to: | SciPost Physics Proceedings |
| Proceedings issue: | The 2nd European AI for Fundamental Physics Conference (EuCAIFCon2025) |
| Ontological classification | |
|---|---|
| Academic field: | Physics |
| Specialties: |
|
Abstract
Physics Beyond the Standard Model (BSM) has yet to be observed at the Large Hadron Collider (LHC), motivating the development of model-agnostic, machine learning-based strategies to probe more regions of the phase space. As many final states have not yet been examined for mass resonances, an accelerated approach to bump-hunting is desirable. BumpNet is a neural network trained to map smoothly falling invariant-mass histogram data to statistical significance values. It provides a unique, automatized approach to mass resonance searches with the capacity to scan hundreds of final states reliably and efficiently.
Current status:
In refereeing
Reports on this Submission
Strengths
1 agnostic algorithm
2 performance demonstrated both on toys and on real analysis
2 performance demonstrated both on toys and on real analysis
Weaknesses
1 performance assume gaussian shape (this is clearly stated)
Report
This is solid work which deserves to be published
Recommendation
Publish (easily meets expectations and criteria for this Journal; among top 50%)