Area 4 meeting: ATLAS+CMS combination conventions

Monday 25 Oct 2021, 18:00 → 20:00 Europe/Zurich

CERN

Andrew Gilbert (Northwestern University (US)), Nicolas Berger (Centre National de la Recherche Scientifique (FR))

Analyses:

• Analyses could be selected from a combination of physics interest, how ready they are for EFT interpretation (e.g. if a RIVET routine is available) and how complex the interpretation could be.

• Could start with 3-4 analyses on each side, ideally with a RIVET routine ready to test the parameterization step. Workspaces could be provided either by each experiment, or built from scratch using a common framework.

• Higgs provides strong constraints, and could be included as at the level of STXS or as ready-made EFT (both have covariance matrices available):

  • The ready-made EFT case is easier to implement, but wouldn't allow to exercise the EFT parameterization code.

  • STXS has a RIVET routine already but it is more complex, especially if many channels are considered. However, the STXS->EFT parameterization would be a useful input to have ready for the future.

• Ensuring that both direct-to-EFT results and reinterpreted results (e.g. STXS) can be used in the same workspaces is also a useful goal of the exercise.

• Important to have analyses from each group (e.g. also Top) to test the impact on the result and also make sure all groups ready for the global combination. Again, the same set doesn't need to be used by both experiments.

• A staging approach could be considered, with a very simple set of analyses as starting point and then building up (not necessarily in the same order for each experiment)

TWiki:

• Important to make sure ATLAS, CMS and theorists are on the same page in terms of framework.

• Important to sort our common conventions for parameter names, signs, numerical factors, etc.

• Some ATLAS additions/suggestion to the TWiki will be provided, but the current version is already very much in line with ATLAS plans.

Parameterization:

• For predictions, should specify more details such as SMEFTsim v2 or 3, etc.

• Useful to have a common code, to ensure consistent parameterizations, and also other codes for cross-checking at least initially. Consistency is not required if the differences can be covered by proper uncertainties, but this may not be fully the case at least initially. The exercise may also be a good test-bed to experiment with different uncertainty schemes.

• Would be useful to make available EFT parameterizations in a way that is easy to share and reuse, especially for common binnings such as STXS (also related uncertainties)

Plans:

• Should identify personpower and target analyses. Could meet again in a few weeks, when this has been established ?

 

 

18:00 → 18:20 Introduction

Speakers: Andrew Gilbert (Northwestern University (US)), Nicolas Berger (Centre National de la Recherche Scientifique (FR))

EFT-Area4-Discussion.pdf