MATRIX HAWAII: PineAPPL interpolation grids with MATRIX Preprint: ZU-TH 41/25 MS-TP-25-08
S. Devoto, T. Ježo, S. Kallweit, C. Schwan · 2025
Read MATRIX HAWAII: PineAPPL interpolation grids with MATRIX Preprint: ZU-TH 41/25 MS-TP-25-08 by S. Devoto, T. Ježo, S. Kallweit, C. Schwan (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-14486.txt.
Introduction
Devoto, Jezo, Kallweit, and Schwan present a software link from twenty twenty-five for particle physics sums. The link joins two tools, PineAPPL and MATRIX, so users can save results as grids. Grids let others redo sums for new input sets without running the full hard sum again. The sums reach a high level of care in two kinds of theory at once. The authors call it the first public tool of its kind at that level.
Such grids help world fits of the proton parts that go into collider sums. Exact sums replace rough scale factors that can skew the fits. The new code and the grids made for the work are posted for free public use. The tool should sharpen later fits and their stated doubt ranges.
Catalog record
| CREATOR | S. Devoto, T. Ježo, S. Kallweit, C. Schwan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 78 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-14486
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
MATRIX HAWAII: PineAPPL interpolation grids with MATRIX Preprint: ZU-TH 41/25 MS-TP-25-08 S. Devoto, T. Ježo, S. Kallweit, C. Schwan Abstract We present an interface between PineAPPL and Matrix , which allows fully differential cross sections to be calculated in the form of interpolation grids, accurate at next-to-next-to-leading order (NNLO) in QCD and next-to-leading order in electroweak (EW) theory. This interface is the first publicly available tool to calculate interpolation grids at NNLO QCD accuracy for a wide set of processes. Interpolation grids provide the functionality to compute predictions for arbitrary parton distribution functions (PDFs) as well as PDF uncertainties without the need to repeat the actual calculation. Another important application of the these grids is to perform global analyses of PDFs using exact NNLO calculations instead of K K -factors, which have several drawbacks. This exact treatment of NNLO corrections is also an important prerequisite for fitting PDFs at next-to-next-to-next-to-leading order level with reliable uncertainties. The new version of the Matrix code interfaced to PineAPPL , as well as the grids produced for this publication, are available on the Matrix website and on PloughShare , respectively. 1 Introduction As we near the high-luminosity phase of the Large Hadron Collider (LHC), we approach an unprecedented level of precision and accuracy for experimental measurements of all standard model (SM) processes. To learn the most from data–theory comparisons, matching this level of precision in the corresponding predictions is required. Especially for well-known processes like Drell–Yan production the experimental uncertainties already rival those from theory predictions. In the absence of any significant tension between theory and experiments, the measurements and corresponding predictions are used to improve knowledge of proton structure encoded in parton distribution functions (PDFs). These are a key ingredient for all other predictions for hadronic collisions, and as we strive for unprecedented precision, the refinement of PDFs becomes paramount. Achieving higher precision makes PDF determinations more computationally costly, but so-called interpolation grids Kluge:2006xs ; Carli:2010rw ; Ball:2010de ; Carrazza:2020gss mitigate this computational challenge to a large extent. In an interpolation grid predictions are stored in a PDF-independent way, so that the most time-consuming part of the computation is only performed once and convolutions with arbitrary PDFs can be performed very quickly a-posteriori. The use of interpolation grids in global analyses of PDFs is by now widespread at next-to-leading order (NLO) QCD and is getting adapted Ball:2011mu ; ATLAS:2018owm ; H1:2021xxi as soon as they are available at next-to-next-to-leading order (NNLO) QCD Britzger:2019kkb ; Britzger:2022lbf ; Czakon:2017dip ; Czakon:2019yrx ; Cruz-Martinez:2025ffa . However, the availability of grids at NNLO QCD is currently limited. Moreover, there is currently no publicly available program that is able to generate NNLO QCD interpolation grids for LHC processes. The goal of this publication is to address this issue by providing a tool capable of producing grids based on a combination of Matrix Grazzini:2017mhc and PineAPPL Carrazza:2020gss ; christopher_schwan_2024_10698034 , named Matrix Hawaii 1 1 1 Matrix Hawaii is the abbreviation of “ Munich – the MUlti-chaNnel Integrator at swiss (CH) precision – Automates qT-subtraction and Resummation to Integrate X-sections, Handling Automation With Additional pIneappl – Pineappl Is Not an Extension of APPLgrid – Interpolation-grids” . Matrix provides predictions for a set of processes that are indispensable for realistic PDF determinations including Drell–Yan off-shell Z / W -boson production and t\bar{t} production up to NNLO QCD+NLO EW Grazzini:2019jkl and NNLO QCD Catani:2019iny ; Catani:2019hip , respectively. Complementary to that, PineAPPL provides the means to grid those predictions, notably for the first time also implementing a way to combine QCD and EW corrections. Providing this tool 2 2 2 This idea was already explored in Ref. Garzelli:2023rvx in a private extension of Matrix , limited to t\bar{t} production and without a generalized extrapolation procedure (see Section 3.2 ). is a crucial step towards making precise predictions in the form of interpolation grids available to a broader community. In most modern PDF extractions at NNLO, the grids are often missing and therefore a K -factor approximation is applied instead, in which the NNLO QCD predictions are obtained by multiplying the NLO QCD predictions with K -factors, defined as the ratio of NNLO to NLO. Whereas these K -factors are usually computed on a bin-by-bin basis, they are assumed to be independent of both the PDFs with which they were generated and the partonic channels. The reliability of this approximation can only be assessed by comparing a fit that uses K -factors to the same fit that instead uses full NNLO calculations. In the absence of such fit, we design a proxy to gain some insight on the impact of using full NNLO calculations instead of relying on K -factors both on the central PDFs and the uncertainty bands. This paper is organized as follows: in Section 2 we briefly describe Matrix and PineAPPL , and detail the new interface. In Section 3 we validate our implementation by producing grids for a small selection of LHC measurements, and comparing them to the results provided by Matrix directly. In Section 4 we showcase possible applications of our interface, including an investigation of the difference between using full NNLO QCD predictions and K -factors in PDF fits. In Section 5 we present our conclusions. Finally, in Appendix A we describe the installation of this toolchain and how to use it, and in Appendix B we describe technical aspects of the implementation that allow us to control the power corrections at the level of interpolation grids. 2 PineAPPL interface to Matrix An interpolation grid is a representation of a (multi-)differential distribution independent of the PDFs and the strong coupling, corresponding to a binned histogram produced by a Monte Carlo (MC) integrator. The advantage of such a representation is that the time-consuming integration of the matrix elements needs to be performed only once; after their generation the grids can be convolved with any PDF set. This convolution with PDFs is an operation that is very fast, usually taking only a fraction of a second. Building on this basic operation, interpolation grids offer further fast operations and applications: • PDF uncertainties . Interpolation grids naturally support the calculation of PDF uncertainties. Without grids, the calculation of PDF uncertainties on-the-fly in NNLO predictions may be available Campbell:2019dru ; Monni:2019whf , but still poses a significant computational challenge due to memory limitations and the speed of IO operations. • Channel sizes . Since interpolation grids store the results for each partonic channel separately, one can study the size of each channel and how it changes in different bins for different PDF sets. This allows us to better understand the process itself and the differences due to the choice of a PDF set. • PDF fits . Since interpolation grids are independent of PDFs, they are an important ingredient of PDF determinations. As soon as a prediction for a measurement is available as an interpolation grid, it can be relatively easily included in a PDF fit. Yet another advantage of interpolation grids is that, once they are generated, the corresponding MC integrators are no longer needed to perform these operations. Therefore, they can be understood as an efficient exchange format, between prediction providers and (possibly different) users, making grids the