Inductive Simulation of Calorimeter Showers with Normalizing Flows

Abstract

Simulating particle detector response is the single most expensive step in the Large Hadron Collider computational pipeline. Recently it was shown that normalizing flows can accelerate this process while achieving unprecedented levels of accuracy, but scaling this approach up to higher resolutions relevant for future detector upgrades leads to prohibitive memory constraints. To overcome this problem, we introduce Inductive CaloFlow (iCaloFlow), a framework for fast detector simulation based on an inductive series of normalizing flows trained on the pattern of energy depositions in pairs of consecutive calorimeter layers. We further use a teacher-student distillation to increase sampling speed without loss of expressivity. As we demonstrate with Datasets 2 and 3 of the CaloChallenge2022, iCaloFlow can realize the potential of normalizing flows in performing fast, high-fidelity simulation on detector geometries that are ~ 10 - 100 times higher granularity than previously considered.

arXiv comment19 pages, 15 figures; v2: title changed, matches published version

References in corpus (15)

• Array Programming with NumPy
• Normalizing Flows: An Introduction and Review of Current Methods
• CaloGAN: Simulating 3D High Energy Particle Showers in Multi-Layer Electromagnetic Calorimeters with Generative Adversarial Networks
• Accelerating Science with Generative Adversarial Networks: An Application to 3D Particle Showers in Multi-Layer Calorimeters
• Getting High: High Fidelity Simulation of High Granularity Calorimeters with High Speed
• Precise simulation of electromagnetic calorimeter showers using a Wasserstein Generative Adversarial Network
• AtlFast3: the next generation of fast simulation in ATLAS
• Calorimetry with Deep Learning: Particle Simulation and Reconstruction for Collider Physics
• Score-based Generative Models for Calorimeter Shower Simulation
• CaloFlow: Fast and Accurate Generation of Calorimeter Showers with Normalizing Flows
• Decoding Photons: Physics in the Latent Space of a BIB-AE Generative Network
• Controlling Physical Attributes in GAN-Accelerated Simulation of Electromagnetic Calorimeters
• CaloClouds: Fast Geometry-Independent Highly-Granular Calorimeter Simulation
• L2LFlows: Generating High-Fidelity 3D Calorimeter Images
• DCTRGAN: Improving the Precision of Generative Models with Reweighting

Cited by in corpus (22)

• The MadNIS Reloaded
• High-dimensional and Permutation Invariant Anomaly Detection
• CaloDREAM -- Detector Response Emulation via Attentive flow Matching
• Precision-Machine Learning for the Matrix Element Method
• Normalizing Flows for High-Dimensional Detector Simulations
• Calorimeter shower superresolution
• Differentiable MadNIS-Lite
• Convolutional L2LFlows: Generating Accurate Showers in Highly Granular Calorimeters Using Convolutional Normalizing Flows
• CaloGraph: Graph-based diffusion model for fast shower generation in calorimeters with irregular geometry
• Accurate Surrogate Amplitudes with Calibrated Uncertainties
• CaloChallenge 2022: A Community Challenge for Fast Calorimeter Simulation
• Anomaly detection with flow-based fast calorimeter simulators
• Unsupervised and lightly supervised learning in particle physics
• Unifying Simulation and Inference with Normalizing Flows
• Extrapolating Jet Radiation with Autoregressive Transformers
• BitHEP -- The Limits of Low-Precision ML in HEP
• Amplitude Uncertainties Everywhere All at Once
• Observable Optimization for Precision Theory: Machine Learning Energy Correlators
• CaloHadronic: a diffusion model for the generation of hadronic showers
• Ultra Fast Calorimeter Simulation with Generative Machine Learning on FPGAs
• ParaFlow: fast calorimeter simulations parameterized in upstream material configurations
• CaloClouds3: Ultra-Fast Geometry-Independent Highly-Granular Calorimeter Simulation