AG-2024.05-1328·hep-ex·cross-listed: hep-ph
Advancing Set-Conditional Set Generation: Diffusion Models for Fast Simulation of Reconstructed Particles
Authors
- Dmitrii Kobylianskii
- Nathalie Soybelman
- Nilotpal Kakati
- Etienne Dreyer
- Benjamin Nachman
- Eilam Gross
Abstract
The computational intensity of detector simulation and event reconstruction poses a significant difficulty for data analysis in collider experiments. This challenge inspires the continued development of machine learning techniques to serve as efficient surrogate models. We propose a fast emulation approach that combines simulation and reconstruction. In other words, a neural network generates a set of reconstructed objects conditioned on input particle sets. To make this possible, we advance set-conditional set generation with diffusion models. Using a realistic, generic, and public detector simulation and reconstruction package (COCOA), we show how diffusion models can accurately model the complex spectrum of reconstructed particles inside jets.
Submitted
16 May 20242 years ago
Version
v1
License
CC-BY-4.0
DOI
10.48550/arXiv.2405.10106
Summary
Researchers use diffusion models—a type of generative AI—to rapidly simulate what particle detectors see after collisions, bypassing slow traditional simulations by directly generating detector outputs from input particles.
- Diffusion models can generate sets of particles conditioned on other sets, a capability previously underexplored for physics simulation.
- The approach combines simulation and reconstruction into one fast neural-network step, potentially accelerating data analysis at collider experiments like the LHC by orders of magnitude.
- The method was validated on realistic detector data (COCOA), showing it can accurately capture the messy spectrum of particles reconstructed inside jets—a key test of practical utility.
curious · generated by claude-haiku-4-5
Chat with this PDF
Ask questions, probe assumptions, request a plain-English summary. Answers cite sections from the preprint itself.
Community
Questions and answers about this paper from other readers. No formal peer review — just a place to think out loud.