Publications (30)
How to GAN away Detector Effects
Marco Bellagente, Anja Butter, Gregor Kasieczka +2
LHC analyses directly comparing data and simulated events bear the danger of using first-principle predictions only as a black-box part of event simulation. We show how simulations…
Interpreting Parton Distributions with Shapley Values
Raphaël Bonnet-Guerrini, Stefano Carrazza, Stefano Forte +3
We show that Shapley values can be used to trace how individual parton distributions (PDFs) shape the theory predictions for high-energy observables computed from them. This provid…
ELSA -- Enhanced latent spaces for improved collider simulations
Benjamin Nachman, Ramon Winterhalder
Simulations play a key role for inference in collider physics. We explore various approaches for enhancing the precision of simulations using machine learning, including interventi…
Precision-Machine Learning for the Matrix Element Method
Theo Heimel, Nathan Huetsch, Ramon Winterhalder +2
The matrix element method is the LHC inference method of choice for limited statistics. We present a dedicated machine learning framework, based on efficient phase-space integratio…
FASTColor -- Full-color Amplitude Surrogate Toolkit for QCD
Javier Mariño Villadamigo, Rikkert Frederix, Tilman Plehn +2
High-multiplicity events remain a bottleneck for LHC simulations due to their computational cost. We present a ML-surrogate approach to accelerate matrix element reweighting from l…
How to GAN LHC Events
Anja Butter, Tilman Plehn, Ramon Winterhalder
Event generation for the LHC can be supplemented by generative adversarial networks, which generate physical events and avoid highly inefficient event unweighting. For top pair pro…
MadSpace -- Event Generation for the Era of GPUs and ML
Theo Heimel, Olivier Mattelaer, Ramon Winterhalder
MadSpace is a new modular phase-space and event-generation library written in C++ with native GPU support via CUDA and HIP. It provides a unified compute-graph-based framework for…
Modern Machine Learning for LHC Physicists
Tilman Plehn, Anja Butter, Barry Dillon +3
Depending on the point of view, modern machine learning is either providing an unprecedented boost to the numerical methods of particle physics, or it is transforming the way we do…
How to GAN Event Unweighting
Mathias Backes, Anja Butter, Tilman Plehn +1
Event generation with neural networks has seen significant progress recently. The big open question is still how such new methods will accelerate LHC simulations to the level requi…
The Monte Carlo Ecosystem in High-Energy Physics: A Primer
Melissa van Beekveld, Enrico Bothmann, Andy Buckley +3
Monte Carlo event generators are the central interface between theoretical calculations and experimental measurements in collider physics. Over several decades, a comprehensive and…
MadNIS -- Neural Multi-Channel Importance Sampling
Theo Heimel, Ramon Winterhalder, Anja Butter +5
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 normal…
Invertible Networks or Partons to Detector and Back Again
Marco Bellagente, Anja Butter, Gregor Kasieczka +5
For simulations where the forward and the inverse directions have a physics meaning, invertible neural networks are especially useful. A conditional INN can invert a detector simul…
The MadNIS Reloaded
Theo Heimel, Nathan Huetsch, Fabio Maltoni +3
In pursuit of precise and fast theory predictions for the LHC, we present an implementation of the MadNIS method in the MadGraph event generator. A series of improvements in MadNIS…
Full and approximated NLO predictions for like-sign W-boson scattering at the LHC
Stefan Dittmaier, Christopher Schwan, Ramon Winterhalder
We report on a recent calculation of next-to-leading-order (NLO) QCD and electroweak corrections to like-sign W-boson scattering at the Large Hadron Collider, including all partoni…
Amplitude Surrogates for Multi-Jet Processes
Luca Beccatini, Fabio Maltoni, Olivier Mattelaer +1
Accurate and efficient amplitude predictions are essential for precision studies of multi-jet processes at the LHC. We introduce a novel neural network architecture that predicts m…
The Physics Behind ML-based Quark-Gluon Taggers
Sophia Vent, Ramon Winterhalder, Tilman Plehn
Jet taggers provide an ideal testbed for applying explainability techniques to powerful ML tools. For theoretically and experimentally challenging quark-gluon tagging, we first ide…
BitHEP -- The Limits of Low-Precision ML in HEP
Claudius Krause, Daohan Wang, Ramon Winterhalder
The increasing complexity of modern neural network architectures demands fast and memory-efficient implementations to mitigate computational bottlenecks. In this work, we evaluate…
Ephemeral Learning -- Augmenting Triggers with Online-Trained Normalizing Flows
Anja Butter, Sascha Diefenbacher, Gregor Kasieczka +4
The large data rates at the LHC require an online trigger system to select relevant collisions. Rather than compressing individual events, we propose to compress an entire data set…
Neural Control Variates at LO and NLO
Theo Heimel, Tilman Plehn, Rebecca Revelli +2
We employ neural control variates to minimize the range of event weights and avoid negative weights for phase-space integration and event generation. A signed control variate, buil…
Accurate Surrogate Amplitudes with Calibrated Uncertainties
Henning Bahl, Nina Elmer, Luigi Favaro +3
Neural networks for LHC physics have to be accurate, reliable, and controlled. Using neural surrogates for the prediction of loop amplitudes as a use case, we first show how activa…
Latent Space Refinement for Deep Generative Models
Ramon Winterhalder, Marco Bellagente, Benjamin Nachman
Deep generative models are becoming widely used across science and industry for a variety of purposes. A common challenge is achieving a precise implicit or explicit representation…
Like-Sign W-Boson Scattering at the LHC -- Approximations and Full Next-to-Leading-Order Predictions
Stefan Dittmaier, Philipp Maierhöfer, Christopher Schwan +1
We present a new calculation of next-to-leading-order corrections of the strong and electroweak interactions to like-sign W-boson scattering at the Large Hadron Collider, implement…
Machine Learning and LHC Event Generation
Anja Butter, Tilman Plehn, Steffen Schumann +48
First-principle simulations are at the heart of the high-energy physics research program. They link the vast data output of multi-purpose detectors with fundamental theory predicti…
MadNIS at NLO
Giovanni De Crescenzo, Javier Mariño Villadamigo, Nina Elmer +4
We combine fast amplitude surrogates with neural importance sampling to accelerate NLO calculations. For virtual corrections, a learned ratio to the Born matrix element with calibr…
Amplitude Uncertainties Everywhere All at Once
Henning Bahl, Nina Elmer, Tilman Plehn +1
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 an…
How to GAN Event Subtraction
Anja Butter, Tilman Plehn, Ramon Winterhalder
Subtracting event samples is a common task in LHC simulation and analysis, and standard solutions tend to be inefficient. We employ generative adversarial networks to produce new e…
Differentiable MadNIS-Lite
Theo Heimel, Olivier Mattelaer, Tilman Plehn +1
Differentiable programming opens exciting new avenues in particle physics, also affecting future event generators. These new techniques boost the performance of current and planned…
Presenting Unbinned Differential Cross Section Results
Miguel Arratia, Anja Butter, Mario Campanelli +10
Machine learning tools have empowered a qualitatively new way to perform differential cross section measurements whereby the data are unbinned, possibly in many dimensions. Unbinne…
Uncertainty in Physics and AI: Taxonomy, Quantification, and Validation
Manuel HauÃmann, Ramon Winterhalder, Maria Ubiali
Reliable uncertainty quantification is essential for the use of machine learning in physics, where scientific discoveries depend on validated probabilistic statements. We provide a…
Targeting Multi-Loop Integrals with Neural Networks
Ramon Winterhalder, Vitaly Magerya, Emilio Villa +5
Numerical evaluations of Feynman integrals often proceed via a deformation of the integration contour into the complex plane. While valid contours are easy to construct, the numeri…