Compressing PDF sets using generative adversarial networks

Carrazza, Stefano; Cruz-Martinez, Juan; Rabemananjara, Tanjona R.

doi:10.1140/epjc/s10052-021-09338-8

Cited by 18 publications

(21 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, there are many simulation-related questions in fundamental physics, where AImethods allow us to make significant progress. Examples going beyond immediate applications to event generation include symbolic regression [130], sample and data compression [57,131], detection of symmetries [132][133][134][135], and many other fascinating new ideas and concepts.…”

Section: Discussionmentioning

confidence: 99%

“…The successful implementation of this idea has led to the NNPDF family of proton PDF determinations [46,[48][49][50] as well as to variants in the context of polarised PDF [51] and nuclear PDF [52,53] global analyses. The current implementation frontier, which has led to the recent NNPDF4.0 determination, involves a suite of contemporary machine learning methods and tools, specifically cross-validation to avoid overtraining, hyperoptimization [54] combined with K-folding for the automatic selection of the methodology, feature scaling of the input for the optimization of the neural networks used as basic underlying model [55], and GAN-enhanced compression for final efficient delivery [56,57].…”

Section: Parton Distribution Functionsmentioning

confidence: 99%

See 1 more Smart Citation

Machine Learning and LHC Event Generation

Butter¹,

Plehn²,

Schumann³

et al. 2022

Preprint

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

Section: Discussionmentioning

confidence: 99%

Section: Parton Distribution Functionsmentioning

confidence: 99%

Machine Learning and LHC Event Generation

Butter¹,

Plehn²,

Schumann³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…However, it usually means encoding a given data point into a more storage-efficient representation. This is different from our proposal which aims to encode the full underlying distribution in the network parameters [28] and to represent all aspects of the LHC training data in the generative network. Modulo limits of expressivity, the online-trained network output can, in principle, replace the training data completely.…”

Section: Introductionmentioning

confidence: 87%

Ephemeral Learning -- Augmenting Triggers with Online-Trained Normalizing Flows

Butter,

Diefenbacher,

Kasieczka

et al. 2022

Preprint

View full text Add to dashboard Cite

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 at once. We use a normalizing flow as a deep generative model to learn the probability density of the data online. The events are then represented by the generative neural network and can be inspected offline for anomalies or used for other analysis purposes. We demonstrate our new approach for a toy model and a correlation-enhanced bump hunt.

show abstract

“…More details about the Monte Carlo compression strategy adopted in this work can be found in App. A.1, with technical information about the algorithmic settings following [167].…”

Section: Monte Carlo Compressionmentioning

confidence: 99%

The PDF4LHC21 combination of global PDF fits for the LHC Run III

Ball,

Butterworth,

Cooper-Sarkar

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

A precise knowledge of the quark and gluon structure of the proton, encoded by the parton distribution functions (PDFs), is of paramount importance for the interpretation of high-energy processes at present and future lepton-hadron and hadron-hadron colliders. Motivated by recent progress in the PDF determinations carried out by the CT, MSHT, and NNPDF groups, we present an updated combination of global PDF fits: PDF4LHC21. It is based on the Monte Carlo combination of the CT18, MSHT20, and NNPDF3.1 sets followed by either its Hessian reduction or its replica compression. Extensive benchmark studies are carried out in order to disentangle the origin of the differences between the three global PDF sets. In particular, dedicated fits based on almost identical theory settings and input datasets are performed by the three groups, highlighting the role played by the respective fitting methodologies. We compare the new PDF4LHC21 combination with its predecessor, PDF4LHC15, demonstrating their good overall consistency and a modest reduction of PDF uncertainties for key LHC processes such as electroweak gauge boson production and Higgs boson production in gluon fusion. We study the phenomenological implications of PDF4LHC21 for a representative selection of inclusive, fiducial, and differential cross sections at the LHC. The PDF4LHC21 combination is made available via the LHAPDF library and provides a robust, user-friendly, and efficient method to estimate the PDF uncertainties associated to theoretical calculations for the upcoming Run III of the LHC and beyond.

show abstract

Compressing PDF sets using generative adversarial networks

Cited by 18 publications

References 28 publications

Machine Learning and LHC Event Generation

Machine Learning and LHC Event Generation

Ephemeral Learning -- Augmenting Triggers with Online-Trained Normalizing Flows

The PDF4LHC21 combination of global PDF fits for the LHC Run III

Contact Info

Product

Resources

About