Welcome to the High Energy Physics Statistics Serialization Standard.
The project is in an early stage and currently resembles more of a working group and not yet a standard.
Everything, including naming, folder structure etc is under construction and open to discussion
The HEP Statistics Serialization Standard (HS3) defines standards of different statistical procedures and results used in High Energy Physics (HEP) in terms of human-and machine readable representations. Different versions are defined with specifications and semantics that are acknowledged by a committee.
View the current draft of the HS3 standard here: https://www.overleaf.com/read/ywzfwjhwvqrv
Corresponding implementations to check the validity of files are also provided, at a best effort basis. More information can be found in the corresponding projects.
There are two main motivation to have a code-independent representation of the likelihood and inference results.
-
It allows to publish the likelihood; a long-term goal in High Energy Physics experiments.
-
A framework- and language independent representation allows to use different frameworks interchangeably; at least for reasonable complicated cases. It removes the dependency on code and reduces the need for maintenance of legacy projects.
HS3 standardizes the machine- and human-readable serialization of all components involved in model fitting as used in High Energy Physics. This includes the definition of the model, the data as well as the loss function. The aim is to provide on one hand a language and framework agnostic, machine-reabable and on the other hand a human-readable, publishable and preservable representation. This would allow to interchange the fitting framework and decouple the reproducibility from the frameworks lifetime.
The focus of this project is not on the long-term storage of large data, which may be needed for an actual implementation, but rather to define a common serialization format.
TODO: more description
Until 1.x, the standard is considered unstable and may introduce backwards incompatible changes.
After that, backwords compatibility is guaranteed to be maintained inside the major releases (1.x, 2.x) and should be kept, if possible, also between the major releases, but is not guaranteed in favor of a cleaner standard.
In order to submit ideas, proposals and examples, you can either start a discussion using issues or add the document(s), if you have some, in the proposals (draft or pending) folder and create a PR to discussion them.
If you are interest to become part of the core committee, please open an issue. Anyone is allowed to join.
To stay updated, you can sign up for the e-group: hep-statistics-serialization-standard@cern.ch
A model fitting library in pure Python. It's focus is on customizability and strong model building.
Pure-Python implementation of statistical model for multi-bin histogram-based analysis and its interval estimation; HistFactory in Python.
The Toolkit for Data Modeling with ROOT (RooFit) is a package that allows for modeling probability distributions in a compact and abstract way.
The Bayesian Analysis Toolkit in Julia.
The ATLAS Collaboration has published measurements in this format:
- Search for same-charge top-quark pair production in
$pp$ collisions at$\sqrt{s}=13$ TeV with the ATLAS detector (HEPdata) - Search for heavy right-handed Majorana neutrinos in the decay of top quarks produced in proton proton collisions at
$\sqrt{s}=13$ TeV with the ATLAS detector (HEPdata) - Observation of
$t\bar{t}$ production in the lepton+jets and dilepton channels in$p$ +Pb collisions at$\sqrt{s_{NN}}=8.16$ TeV with the ATLAS detector (HEPdata) - Measurements of inclusive and differential cross-sections of
$t\bar{t}\gamma$ production in$pp$ collisions at$\sqrt{s}=13$ TeV with the ATLAS detector (HEPdata - Measurement of the
$t\bar{t}$ cross section and its ratio to the$Z$ production cross section using$pp$ collisions at$\sqrt{s}=13.6$ TeV with the ATLAS detector (HEPdata) - Inclusive and differential cross-section measurements of
$t\bar{t}Z$ production in$pp$ collisions at$\sqrt{s}=13.6$ TeV with the ATLAS detector, including EFT and spin-correlation interpretations (HEPdata) - Measurement of the charge asymmetry in top-quark pair production in association with a photon with the ATLAS experiment (HEPdata)
The comittee is responsible for the acceptance and denial of new proposals and has to approve a new version of the standard.