CaloFlow: Fast and Accurate Generation of Calorimeter Showers with Normalizing Flows

Abstract

We introduce CaloFlow, a fast detector simulation framework based on normalizing flows. For the first time, we demonstrate that normalizing flows can reproduce many-channel calorimeter showers with extremely high fidelity, providing a fresh alternative to computationally expensive GEANT4 simulations, as well as other state-of-the-art fast simulation frameworks based on GANs and VAEs. Besides the usual histograms of physical features and images of calorimeter showers, we introduce a new metric for judging the quality of generative modeling: the performance of a classifier trained to differentiate real from generated images. We show that GAN-generated images can be identified by the classifier with nearly 100% accuracy, while images generated from CaloFlow are better able to fool the classifier. More broadly, normalizing flows offer several advantages compared to other state-of-the-art approaches (GANs and VAEs), including: tractable likelihoods; stable and convergent training; and principled model selection. Normalizing flows also provide a bijective mapping between data and the latent space, which could have other applications beyond simulation, for example, to detector unfolding.

33 pages, 19 figures, 5 tables; v2: improved handling of datasets, conclusions unchanged; v3: matches accepted version

References in corpus (11)

• Array Programming with NumPy
• Understanding Event-Generation Networks via Uncertainties
• Decoding Photons: Physics in the Latent Space of a BIB-AE Generative Network
• Fast simulation of muons produced at the SHiP experiment using Generative Adversarial Networks
• Phase Space Sampling and Inference from Weighted Events with Autoregressive Flows
• Cherenkov Detectors Fast Simulation Using Neural Networks
• Exhaustive Neural Importance Sampling applied to Monte Carlo event generation
• Machine Learning Templates for QCD Factorization in the Search for Physics Beyond the Standard Model
• Compressing PDF sets using generative adversarial networks
• The electromagnetic decays of $X(3823)$ as the ${\psi }_2(1^3{D}_2)$ state and its radial excited states
• A Data-driven Event Generator for Hadron Colliders using Wasserstein Generative Adversarial Network

Cited by in corpus (58)

• Quantum Computing for High-Energy Physics: State of the Art and Challenges. Summary of the QC4HEP Working Group
• Event Generators for High-Energy Physics Experiments
• Fast Point Cloud Generation with Diffusion Models in High Energy Physics
• CaloClouds: Fast Geometry-Independent Highly-Granular Calorimeter Simulation
• EPiC-GAN: Equivariant Point Cloud Generation for Particle Jets
• PC-JeDi: Diffusion for Particle Cloud Generation in High Energy Physics
• MadNIS -- Neural Multi-Channel Importance Sampling
• L2LFlows: Generating High-Fidelity 3D Calorimeter Images
• How to Understand Limitations of Generative Networks
• Two Invertible Networks for the Matrix Element Method
• CaloScore v2: Single-shot Calorimeter Shower Simulation with Diffusion Models
• Deep generative models for fast photon shower simulation in ATLAS
• Inductive Simulation of Calorimeter Showers with Normalizing Flows
• The MadNIS Reloaded
• OmniJet-$\alpha$: The first cross-task foundation model for particle physics
• PC-Droid: Faster diffusion and improved quality for particle cloud generation
• CaloFlow for CaloChallenge Dataset 1
• High-dimensional and Permutation Invariant Anomaly Detection
• CaloDREAM -- Detector Response Emulation via Attentive flow Matching
• Precision-Machine Learning for the Matrix Element Method
• ELSA -- Enhanced latent spaces for improved collider simulations
• Normalizing Flows for High-Dimensional Detector Simulations
• Calorimeter shower superresolution
• An implementation of neural simulation-based inference for parameter estimation in ATLAS
• Convolutional L2LFlows: Generating Accurate Showers in Highly Granular Calorimeters Using Convolutional Normalizing Flows
• Differentiable MadNIS-Lite
• CaloShowerGAN, a Generative Adversarial Networks model for fast calorimeter shower simulation
• Goodness of fit by Neyman-Pearson testing
• Generalizing to new geometries with Geometry-Aware Autoregressive Models (GAAMs) for fast calorimeter simulation
• End-to-end simulation of particle physics events with Flow Matching and generator Oversampling
• CaloChallenge 2022: A Community Challenge for Fast Calorimeter Simulation
• Anomaly detection with flow-based fast calorimeter simulators
• Accurate Surrogate Amplitudes with Calibrated Uncertainties
• CaloGraph: Graph-based diffusion model for fast shower generation in calorimeters with irregular geometry
• Aspen Open Jets: Unlocking LHC Data for Foundation Models in Particle Physics
• Unsupervised and lightly supervised learning in particle physics
• GalaxyFlow: Upsampling Hydrodynamical Simulations for Realistic Mock Stellar Catalogs
• Advancing Set-Conditional Set Generation: Diffusion Models for Fast Simulation of Reconstructed Particles
• Flow Matching Beyond Kinematics: Generating Jets with Particle-ID and Trajectory Displacement Information
• Parnassus: An Automated Approach to Accurate, Precise, and Fast Detector Simulation and Reconstruction
• Unifying Simulation and Inference with Normalizing Flows
• Light Dark Matter Search with 7.8 Tonne-Year of Ionization-Only Data in XENONnT
• The LHCb ultra-fast simulation option, Lamarr: design and validation
• Extrapolating Jet Radiation with Autoregressive Transformers
• BitHEP -- The Limits of Low-Precision ML in HEP
• Amplitude Uncertainties Everywhere All at Once
• Differentiable Surrogate for Detector Simulation and Design with Diffusion Models
• Choose Your Diffusion: Efficient and flexible ways to accelerate the diffusion model in fast high energy physics simulation
• CaloHadronic: a diffusion model for the generation of hadronic showers
• Reconstructing Sparticle masses at the LHC using Generative Machine Learning
• Observable Optimization for Precision Theory: Machine Learning Energy Correlators
• One flow to correct them all: improving simulations in high-energy physics with a single normalising flow and a switch
• ParaFlow: fast calorimeter simulations parameterized in upstream material configurations
• ScatterPrism: convergence for generative simulation and inverse problems in particle and nuclear physics
• Data-Driven Predictions for Dark Photon and Millicharged Particle Production
• Implicit Quantile Networks For Emulation in Jet Physics
• Ultra-Fast Generation of Air Shower Images for Imaging Air Cherenkov Telescopes using Generative Adversarial Networks
• CaloClouds3: Ultra-Fast Geometry-Independent Highly-Granular Calorimeter Simulation