Machine learning amplitudes for faster event generation

Bishara, Fady; Montull, Marc

doi:10.1103/physrevd.107.l071901

Cited by 11 publications

(11 citation statements)

References 77 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In particular, BDTs have been the workhorse of particle physics for a long time but mostly for performing classification of tiny signals from dominating backgrounds 6 . However, the utility of BDTs as regressors for theoretical estimates of experimental signatures has only been advocated recently 7 and has been shown to achieve impressive accuracy for 2D data.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

High-precision regressors for particle physics

Bishara,

Paul,

2024

Sci Rep

Self Cite

View full text Add to dashboard Cite

Monte Carlo simulations of physics processes at particle colliders like the Large Hadron Collider at CERN take up a major fraction of the computational budget. For some simulations, a single data point takes seconds, minutes, or even hours to compute from first principles. Since the necessary number of data points per simulation is on the order of $$10^9$$ 10 9 –$$10^{12}$$ 10 12 , machine learning regressors can be used in place of physics simulators to significantly reduce this computational burden. However, this task requires high-precision regressors that can deliver data with relative errors of less than 1% or even 0.1% over the entire domain of the function. In this paper, we develop optimal training strategies and tune various machine learning regressors to satisfy the high-precision requirement. We leverage symmetry arguments from particle physics to optimize the performance of the regressors. Inspired by ResNets, we design a Deep Neural Network with skip connections that outperform fully connected Deep Neural Networks. We find that at lower dimensions, boosted decision trees far outperform neural networks while at higher dimensions neural networks perform significantly better. We show that these regressors can speed up simulations by a factor of $$10^3$$ 10 3 –$$10^6$$ 10 6 over the first-principles computations currently used in Monte Carlo simulations. Additionally, using symmetry arguments derived from particle physics, we reduce the number of regressors necessary for each simulation by an order of magnitude. Our work can significantly reduce the training and storage burden of Monte Carlo simulations at current and future collider experiments.

show abstract

Section: Related Workmentioning

confidence: 99%

“…In contrast to prior works 7 , 8 , 24 , 25 , the novelty of our contribution is that we try to attain high precision in the entire domain of the function being regressed with fast and efficient regressors. For that, we compare BDTs and neural networks for functions with 2D, 4D, and 8D input feature spaces.…”

Section: Related Workmentioning

confidence: 99%

High-precision regressors for particle physics

Bishara,

Paul,

2024

Sci Rep

Self Cite

View full text Add to dashboard Cite

show abstract

“…Perturbative precision calculations use interpolation methods to reduce the evaluation time for expensive loop amplitudes, defining a task where appropriately designed neural networks can be expected to outperform standard methods [26][27][28][29][30]. The challenge in NN-based surrogate models for integrands and amplitudes is to ensure that all relevant features are indeed encoded in the network at sufficient precision and to establish a reliable uncertainty treatment of the network training.…”

Section: Scattering Amplitudesmentioning

confidence: 99%

Machine learning and LHC event generation

et al. 2023

View full text Add to dashboard Cite

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 predictions and interpretation. This review illustrates a wide range of applications of modern machine learning to event generation and simulation-based inference, including conceptional developments driven by the specific requirements of particle physics. New ideas and tools developed at the interface of particle physics and machine learning will improve the speed and precision of forward simulations, handle the complexity of collision data, and enhance inference as an inverse simulation problem.

show abstract

“…Underlying techniques include generative adversarial networks (GANs) [1][2][3], variational autoencoders (VAEs) [4,5], normalizing flows [6][7][8][9][10], and their invertible network (INN) variant [11][12][13]. As part of the standard LHC event generation chain [14], modern neural networks can be applied to the full range of phase space integration [15,16], phase space sampling [17][18][19][20], amplitude computations [21,22], event subtraction [23], event unweighting [24,25], parton showering [26][27][28][29][30], or super-resolution enhancement [31,32]. Conceptionally new developments are, for instance, based on fully NN-based event generators [33][34][35][36][37] or detector simulations [38][39][40][41][42][43][44][45][4...…”

Section: Introductionmentioning

confidence: 99%

Generative networks for precision enthusiasts

Butter¹,

Heimel²,

Hummerich³

et al. 2023

SciPost Phys.

View full text Add to dashboard Cite

Generative networks are opening new avenues in fast event generation for the LHC. We show how generative flow networks can reach percent-level precision for kinematic distributions, how they can be trained jointly with a discriminator, and how this discriminator improves the generation. Our joint training relies on a novel coupling of the two networks which does not require a Nash equilibrium. We then estimate the generation uncertainties through a Bayesian network setup and through conditional data augmentation, while the discriminator ensures that there are no systematic inconsistencies compared to the training data.

show abstract

Machine learning amplitudes for faster event generation

Cited by 11 publications

References 77 publications

High-precision regressors for particle physics

High-precision regressors for particle physics

Machine learning and LHC event generation

Generative networks for precision enthusiasts

Contact Info

Product

Resources

About