Generative Diffusion Models for Fast Simulations of Particle Collisions at CERN

Mikołaj Kita,Jan Dubiński,Przemysław Rokita,Kamil Deja

2024-06-05

Abstract:In High Energy Physics simulations play a crucial role in unraveling the complexities of particle collision experiments within CERN's Large Hadron Collider. Machine learning simulation methods have garnered attention as promising alternatives to traditional approaches. While existing methods mainly employ Variational Autoencoders (VAEs) or Generative Adversarial Networks (GANs), recent advancements highlight the efficacy of diffusion models as state-of-the-art generative machine learning methods. We present the first simulation for Zero Degree Calorimeter (ZDC) at the ALICE experiment based on diffusion models, achieving the highest fidelity compared to existing baselines. We perform an analysis of trade-offs between generation times and the simulation quality. The results indicate a significant potential of latent diffusion model due to its rapid generation time.

Data Analysis, Statistics and Probability,Computer Vision and Pattern Recognition,High Energy Physics - Experiment

What problem does this paper attempt to address?

The paper aims to address the issue of particle collision experiment simulation in High Energy Physics (HEP), specifically focusing on the Zero Degree Calorimeter (ZDC) in the Large Hadron Collider (LHC) at the European Organization for Nuclear Research (CERN) for fast and accurate simulation. Traditionally, these simulations rely on Monte Carlo methods, which, despite producing high-quality results, are computationally expensive. To optimize resource usage, researchers have begun exploring the use of generative machine learning methods to reduce CPU consumption. Specifically, this paper proposes a new method based on diffusion models for ZDC simulation in the ALICE experiment and demonstrates the superior performance of this method in terms of the Wasserstein distance evaluation metric, with the generated image quality surpassing existing baseline methods. Additionally, the paper discusses the trade-off between generation time and simulation quality, noting that the Latent Diffusion Model has significant potential due to its fast generation time but still requires further development to meet the performance standards of other models.

Generative Diffusion Models for Fast Simulations of Particle Collisions at CERN

Deep Generative Models for Proton Zero Degree Calorimeter Simulations in ALICE, CERN

Applying generative neural networks for fast simulations of the ALICE (CERN) experiment

Machine Learning methods for simulating particle response in the Zero Degree Calorimeter at the ALICE experiment, CERN

Particle physics DL-simulation with control over generated data properties

Diffusion model approach to simulating electron-proton scattering events

Advancing Set-Conditional Set Generation: Diffusion Models for Fast Simulation of Reconstructed Particles

A Comprehensive Evaluation of Generative Models in Calorimeter Shower Simulation

Choose Your Diffusion: Efficient and flexible ways to accelerate the diffusion model in fast high energy physics simulation

Effectiveness of denoising diffusion probabilistic models for fast and high-fidelity whole-event simulation in high-energy heavy-ion experiments

CaloGraph: Graph-based diffusion model for fast shower generation in calorimeters with irregular geometry

Fast Point Cloud Generation with Diffusion Models in High Energy Physics

Denoising diffusion models with geometry adaptation for high fidelity calorimeter simulation

Point cloud-based diffusion models for the Electron-Ion Collider

Graph-based diffusion model for fast shower generation in calorimeters with irregular geometry

Getting High: High Fidelity Simulation of High Granularity Calorimeters with High Speed

Fast and accurate simulation of particle detectors using generative adversarial networks

CaloGAN: Simulating 3D High Energy Particle Showers in Multi-Layer Electromagnetic Calorimeters with Generative Adversarial Networks

CaloClouds: Fast Geometry-Independent Highly-Granular Calorimeter Simulation

Systematic Evaluation of Generative Machine Learning Capability to Simulate Distributions of Observables at the Large Hadron Collider