Exploring phase space with Neural Importance Sampling

Bothmann, Enrico; Janßen, Timo; Knobbe, Max; Schmale, Tobias; Schumann, Steffen

doi:10.21468/scipostphys.8.4.069

Cited by 100 publications

(130 citation statements)

References 70 publications

Supporting

Mentioning

130

Contrasting

Order By: Relevance

“…The number of times these amplitudes are evaluated for a given process is determined by the integration method. Recent work, such as that of [10,11], develop novel methods for the integration of scattering cross sections. These methods have the potential to be combined with new techniques for efficient matrix element computation, such as those presented in this study, to provide even greater cost savings when calculating cross-sections and differentials.…”

Section: Discussionmentioning

confidence: 99%

“…Previous attempts have been made to use machine learning tools such as Boosted Decision Trees (BDTs) and neural networks for efficient phase-space sampling and Monte Carlo integration [8,9] with recent work [10,11] focusing on the use of coupling layers [12]. Similarly, work such as that of Otten et al [13] makes use of neural networks for explicit cross-section prediction.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Using neural networks for efficient evaluation of high multiplicity scattering amplitudes

Badger

Bullock

2020

J. High Energ. Phys.

View full text Add to dashboard Cite

Precision theoretical predictions for high multiplicity scattering rely on the evaluation of increasingly complicated scattering amplitudes which come with an extremely high CPU cost. For state-of-the-art processes this can cause technical bottlenecks in the production of fully differential distributions. In this article we explore the possibility of using neural networks to approximate multi-variable scattering amplitudes and provide efficient inputs for Monte Carlo integration. We focus on QCD corrections to e + e − → jets up to one-loop and up to five jets. We demonstrate reliable interpolation when a series of networks are trained to amplitudes that have been divided into sectors defined by their infrared singularity structure. Complete simulations for one-loop distributions show speed improvements of at least an order of magnitude over a standard approach.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Using neural networks for efficient evaluation of high multiplicity scattering amplitudes

Badger

Bullock

2020

J. High Energ. Phys.

View full text Add to dashboard Cite

show abstract

“…INNs are an alternative class of generative networks, based on normalizing flows [38][39][40][41]. In particle physics such normalizing flow networks have proven useful for instance in phase space generation [42], linking integration with generation [43,44], or anomaly detection [45].…”

Section: Introductionmentioning

confidence: 99%

Invertible networks or partons to detector and back again

et al. 2020

View full text Add to dashboard Cite

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 simulation in terms of high-level observables, specifically for ZW production at the LHC. It allows for a per-event statistical interpretation. Next, we allow for a variable number of QCD jets. We unfold detector effects and QCD radiation to a pre-defined hard process, again with a per-event probabilistic interpretation over parton-level phase space.

show abstract

“…These approaches, however, require learning the full phase space density, instead of just the likelihood ratio, which is significantly more complicated than the neural resampling approach presented here. It is also worth mentioning that neural networks and other machine learning techniques have been studied to improve other components of event generation, including parton density modeling [65,66], phase space generation [67][68][69][70][71], matrix element calculations [72,73], and more [74][75][76].…”

Section: Introductionmentioning

confidence: 99%

Neural resampler for Monte Carlo reweighting with preserved uncertainties

Nachman

Thaler

2020

Phys. Rev. D

View full text Add to dashboard Cite

event generators are an essential tool for data analysis in collider physics. To include subleading quantum corrections, these generators often need to produce negative weight events, which leads to statistical dilution of the datasets and downstream computational costs for detector simulation. Building on the recent proposal of a positive resampler method to rebalance weights within histogram bins, we introduce neural resampling: an unbinned approach to Monte Carlo reweighting based on neural networks that scales well to high-dimensional and variable-dimensional phase space. We pay particular attention to preserving the statistical properties of the event sample, such that neural resampling not only maintains the mean value of any observable but also its Monte Carlo uncertainty. This uncertainty preservation scheme is general and can also be applied to binned (non-neural network) resampling. To illustrate our neural resampling approach, we present a case study from the Large Hadron Collider of top quark pair production at next-to-leading order matched to a parton shower.

show abstract

Exploring phase space with Neural Importance Sampling

Cited by 100 publications

References 70 publications

Using neural networks for efficient evaluation of high multiplicity scattering amplitudes

Using neural networks for efficient evaluation of high multiplicity scattering amplitudes

Invertible networks or partons to detector and back again

Neural resampler for Monte Carlo reweighting with preserved uncertainties

Contact Info

Product

Resources

About