Amplitude Uncertainties Everywhere All at Once

2025/08/29 by Bahl, Henning, Elmer, Nina, Plehn, Tilman +1 · 6 citations
#FOS: Physical sciences #High Energy Physics - Phenomenology (hep-ph)

paper · doi:10.48550/arxiv.2509.00155

Abstract

Ultra-fast, precise, and controlled amplitude surrogates are essential for future LHC event generation. First, we investigate the noise reduction and biases of network ensembles and outline a new method to learn well-calibrated systematic uncertainties for them. We also establish evidential regression as a sampling-free method for uncertainty quantification. In a second part, we tackle localized disturbances for amplitude regression and demonstrate that learned uncertainties from Bayesian networks, ensembles, and evidential regression all identify numerical noise or gaps in the training data.

Citations

• Lorentz-Equivariance without Limitations
• Efficient many-jet event generation with Flow Matching
• Accelerating multijet-merged event generation with neural network matrix element surrogates
• Lorentz Local Canonicalization: How to Make Any Network Lorentz-Equivariant
• Sampling NNLO QCD phase space with normalizing flows
• Evidential Deep Learning for Uncertainty Quantification and Out-of-Distribution Detection in Jet Identification using Deep Neural Networks
• Anomalous electroweak physics unraveled via evidential deep learning
• Accurate Surrogate Amplitudes with Calibrated Uncertainties
• Interpolating amplitudes
• Precision calibration of calorimeter signals in the ATLAS experiment using an uncertainty-aware neural network
• A Lorentz-Equivariant Transformer for All of the LHC
• CaloChallenge 2022: A Community Challenge for Fast Calorimeter Simulation
• Advancing Tools for Simulation-Based Inference
• Lorentz-Equivariant Geometric Algebra Transformers for High-Energy Physics
• PINNferring the Hubble Function with Uncertainties
• Accelerating HEP simulations with Neural Importance Sampling
• Refining Fast Calorimeter Simulations with a Schrödinger Bridge
• Generative Machine Learning for Detector Response Modeling with a Conditional Normalizing Flow
• Faithful Heteroscedastic Regression with Neural Networks
• CaloMan: Fast generation of calorimeter showers with density estimation on learned manifolds
• Modern Machine Learning for LHC Physicists
• Loop Amplitudes from Precision Networks
• On the Pitfalls of Heteroscedastic Uncertainty Estimation with Probabilistic Neural Networks
• Hadrons, Better, Faster, Stronger
• CaloFlow II: Even Faster and Still Accurate Generation of Calorimeter Showers with Normalizing Flows
• Repulsive Deep Ensembles are Bayesian
• Multivariate Deep Evidential Regression
• (Machine) Learning amplitudes for faster event generation
• Reliable training and estimation of variance networks
• LHC analysis-specific datasets with Generative Adversarial Networks

Cited by

• Neural Control Variates at LO and NLO
• Economical Jet Taggers -- Equivariant, Slim, and Quantized
• Amplitude Surrogates for Multi-Jet Processes
• Uncertainty in Physics and AI: Taxonomy, Quantification, and Validation
• Forecasting Generative Amplification
• Iterative HOMER with uncertainties

Related

• Loop Amplitudes from Precision Networks
• Accurate Surrogate Amplitudes with Calibrated Uncertainties
• Forecasting Generative Amplification
• Amplitude Surrogates for Multi-Jet Processes
• Event Generator Tuning Incorporating Systematic Uncertainty