How to Understand Limitations of Generative Networks

Abstract

Well-trained classifiers and their complete weight distributions provide us with a well-motivated and practicable method to test generative networks in particle physics. We illustrate their benefits for distribution-shifted jets, calorimeter showers, and reconstruction-level events. In all cases, the classifier weights make for a powerful test of the generative network, identify potential problems in the density estimation, relate them to the underlying physics, and tie in with a comprehensive precision and uncertainty treatment for generative networks.

arXiv comment32 pages, 19 figures

References in corpus (58)

• ParticleNet: Jet Tagging via Particle Clouds
• CaloGAN: Simulating 3D High Energy Particle Showers in Multi-Layer Electromagnetic Calorimeters with Generative Adversarial Networks
• Anomaly Detection with Density Estimation
• Learning Particle Physics by Example: Location-Aware Generative Adversarial Networks for Physics Synthesis
• Accelerating Science with Generative Adversarial Networks: An Application to 3D Particle Showers in Multi-Layer Calorimeters
• Event Generation with Normalizing Flows
• How to GAN LHC Events
• OmniFold: A Method to Simultaneously Unfold All Observables
• Getting High: High Fidelity Simulation of High Granularity Calorimeters with High Speed
• Precise simulation of electromagnetic calorimeter showers using a Wasserstein Generative Adversarial Network
• JUNIPR: a Framework for Unsupervised Machine Learning in Particle Physics
• Machine Learning and LHC Event Generation
• AtlFast3: the next generation of fast simulation in ATLAS
• Classifying Anomalies THrough Outer Density Estimation (CATHODE)
• Calorimetry with Deep Learning: Particle Simulation and Reconstruction for Collider Physics
• i-flow: High-dimensional Integration and Sampling with Normalizing Flows
• Score-based Generative Models for Calorimeter Shower Simulation
• Exploring phase space with Neural Importance Sampling
• Invertible Networks or Partons to Detector and Back Again
• CaloFlow: Fast and Accurate Generation of Calorimeter Showers with Normalizing Flows
• DijetGAN: A Generative-Adversarial Network Approach for the Simulation of QCD Dijet Events at the LHC
• How to GAN away Detector Effects
• Unfolding with Generative Adversarial Networks
• Understanding Event-Generation Networks via Uncertainties
• Accelerating Monte Carlo event generation -- rejection sampling using neural network event-weight estimates
• Decoding Photons: Physics in the Latent Space of a BIB-AE Generative Network
• Generative Networks for Precision Enthusiasts
• GANplifying Event Samples
• A factorisation-aware Matrix element emulator
• Controlling Physical Attributes in GAN-Accelerated Simulation of Electromagnetic Calorimeters
• Fast Point Cloud Generation with Diffusion Models in High Energy Physics
• CaloClouds: Fast Geometry-Independent Highly-Granular Calorimeter Simulation
• Optimising simulations for diphoton production at hadron colliders using amplitude neural networks
• Evaluating generative models in high energy physics
• Phase Space Sampling and Inference from Weighted Events with Autoregressive Flows
• Mapping Machine-Learned Physics into a Human-Readable Space
• Resonant anomaly detection without background sculpting
• EPiC-GAN: Equivariant Point Cloud Generation for Particle Jets
• MadNIS -- Neural Multi-Channel Importance Sampling
• L2LFlows: Generating High-Fidelity 3D Calorimeter Images
• FETA: Flow-Enhanced Transportation for Anomaly Detection
• Improved Neural Network Monte Carlo Simulation
• DCTRGAN: Improving the Precision of Generative Models with Reweighting
• Reweighting a parton shower using a neural network: the final-state case
• Two Invertible Networks for the Matrix Element Method
• Measuring QCD Splittings with Invertible Networks
• CaloMan: Fast generation of calorimeter showers with density estimation on learned manifolds
• Towards a Computer Vision Particle Flow
• Jet Diffusion versus JetGPT -- Modern Networks for the LHC
• One-loop matrix element emulation with factorisation awareness
• How to GAN Event Subtraction
• New directions for surrogate models and differentiable programming for High Energy Physics detector simulation
• End-To-End Latent Variational Diffusion Models for Inverse Problems in High Energy Physics
• ELSA -- Enhanced latent spaces for improved collider simulations
• Generative Machine Learning for Detector Response Modeling with a Conditional Normalizing Flow
• An unfolding method based on conditional Invertible Neural Networks (cINN) using iterative training
• Feature Selection with Distance Correlation
• CURTAINs Flows For Flows: Constructing Unobserved Regions with Maximum Likelihood Estimation

Cited by in corpus (22)

• Deep generative models for fast photon shower simulation in ATLAS
• The MadNIS Reloaded
• Deep Generative Models for Detector Signature Simulation: A Taxonomic Review
• High-dimensional and Permutation Invariant Anomaly Detection
• CaloDREAM -- Detector Response Emulation via Attentive flow Matching
• The Landscape of Unfolding with Machine Learning
• Precision-Machine Learning for the Matrix Element Method
• Normalizing Flows for High-Dimensional Detector Simulations
• Differentiable MadNIS-Lite
• 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
• Calibrating Bayesian Generative Machine Learning for Bayesiamplification
• Aspen Open Jets: Unlocking LHC Data for Foundation Models in Particle Physics
• Jet Classification Using High-Level Features from Anatomy of Top Jets
• Flow Matching Beyond Kinematics: Generating Jets with Particle-ID and Trajectory Displacement Information
• Generative Unfolding with Distribution Mapping
• BitHEP -- The Limits of Low-Precision ML in HEP
• Extrapolating Jet Radiation with Autoregressive Transformers
• How to Unfold Top Decays
• Event generation and consistency tests with sliced Wasserstein distance in high-energy physics
• Implicit Quantile Networks For Emulation in Jet Physics