MadNIS -- Neural Multi-Channel Importance Sampling

Abstract

Theory predictions for the LHC require precise numerical phase-space integration and generation of unweighted events. We combine machine-learned multi-channel weights with a normalizing flow for importance sampling, to improve classical methods for numerical integration. We develop an efficient bi-directional setup based on an invertible network, combining online and buffered training for potentially expensive integrands. We illustrate our method for the Drell-Yan process with an additional narrow resonance.

arXiv comment33 pages, 15 figures, minor fixes to v1

References in corpus (43)

• The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations
• LHAPDF6: parton density access in the LHC precision era
• NICE: Non-linear Independent Components Estimation
• Variational Inference with Normalizing Flows
• Event Generation with Sherpa 2.2
• MadEvent: Automatic Event Generation with MadGraph
• WHIZARD: Simulating Multi-Particle Processes at LHC and ILC
• The Path to Proton Structure at One-Percent Accuracy
• Guided Image Generation with Conditional Invertible Neural Networks
• Weight optimization in multichannel Monte Carlo
• Event Generation with Normalizing Flows
• Adaptive Multidimensional Integration: VEGAS Enhanced
• Vegas Revisited: Adaptive Monte Carlo Integration Beyond Factorization
• Machine Learning and LHC Event Generation
• i-flow: High-dimensional Integration and Sampling with Normalizing Flows
• 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
• LUSIFER: a LUcid approach to SIx-FERmion production
• Understanding Event-Generation Networks via Uncertainties
• Accelerating Monte Carlo event generation -- rejection sampling using neural network event-weight estimates
• Introduction to Monte Carlo methods
• Neural Network-Based Approach to Phase Space Integration
• Generative Networks for Precision Enthusiasts
• A factorisation-aware Matrix element emulator
• Optimising simulations for diphoton production at hadron colliders using amplitude neural networks
• Targeting Multi-Loop Integrals with Neural Networks
• Phase Space Sampling and Inference from Weighted Events with Autoregressive Flows
• Improved Neural Network Monte Carlo Simulation
• Using neural networks for efficient evaluation of high multiplicity scattering amplitudes
• Two Invertible Networks for the Matrix Element Method
• Measuring QCD Splittings with Invertible Networks
• Inference of cosmic-ray source properties by conditional invertible neural networks
• CaloMan: Fast generation of calorimeter showers with density estimation on learned manifolds
• Modern Machine Learning for LHC Physicists
• ν-Flows: Conditional Neutrino Regression
• CaloFlow II: Even Faster and Still Accurate Generation of Calorimeter Showers with Normalizing Flows
• Lectures on LHC Physics
• Parallel Adaptive Monte Carlo Integration with the Event Generator WHIZARD
• CaloFlow for CaloChallenge Dataset 1
• Ephemeral Learning -- Augmenting Triggers with Online-Trained Normalizing Flows
• Multi-variable Integration with a Neural Network
• Differentiable Matrix Elements with MadJax

Cited by in corpus (32)

• How to Understand Limitations of Generative Networks
• The MadNIS Reloaded
• Efficient phase-space generation for hadron collider event simulation
• CaloDREAM -- Detector Response Emulation via Attentive flow Matching
• Precision-Machine Learning for the Matrix Element Method
• A Lorentz-Equivariant Transformer for All of the LHC
• Normalizing Flows for High-Dimensional Detector Simulations
• Calorimeter shower superresolution
• A Portable Parton-Level Event Generator for the High-Luminosity LHC
• Differentiable MadNIS-Lite
• Convolutional L2LFlows: Generating Accurate Showers in Highly Granular Calorimeters Using Convolutional Normalizing Flows
• Optimal, fast, and robust inference of reionization-era cosmology with the 21cmPIE-INN
• Comparison of Affine and Rational Quadratic Spline Coupling and Autoregressive Flows through Robust Statistical Tests
• Accurate Surrogate Amplitudes with Calibrated Uncertainties
• Anomaly detection with flow-based fast calorimeter simulators
• PIPPIN: Generating variable length full events from partons
• Sampling NNLO QCD phase space with normalizing flows
• Accelerating multijet-merged event generation with neural network matrix element surrogates
• Unifying Simulation and Inference with Normalizing Flows
• BitHEP -- The Limits of Low-Precision ML in HEP
• Extrapolating Jet Radiation with Autoregressive Transformers
• Flow Annealed Importance Sampling Bootstrap meets Differentiable Particle Physics
• Amplitude Uncertainties Everywhere All at Once
• Tropical sampling from Feynman measures
• How to Unfold Top Decays
• Amplitude Surrogates for Multi-Jet Processes
• Monte Carlo Event Generators
• Observable Optimization for Precision Theory: Machine Learning Energy Correlators
• On the simulated kinematic distributions of heavy meson decays
• Monte Carlo Event Generators for Future Lepton Colliders
• Data-Driven Predictions for Dark Photon and Millicharged Particle Production
• NNLO QCD corrections to $\gamma \gamma \to Q\bar{Q}$ from Local Unitarity combined with Coulomb resummation and NLO EW effects