OpenMAPP informal teams catch-up

→ Europe/Zurich
Zoom Meeting ID
69450692292
Host
Andy Buckley
Alternative host
Sabine Kraml
Useful links
Join via phone
Zoom URL

Minutes

Reporting inputs needed urgently. Examples:

  • Martin & Tomek + ADL team to upload data from correlation studies to OpenMAPP Zenodo
  • Rivet, Contur, SModelS, ADL analysis-coverage JSONs published
  • HS3 discussions and standard evolution
  • Contur updates to Spey and switchable LLR/CLs pool-optimisation
  • ...

 

Reports:

  • HS3 discussion with Carsten: their plan is to move away from the "histfactory_dist" and toward a sort of "macro" or saved-pattern of feature combinations that enables more efficient saving of HF or other applications' use-cases. So the request is that rather than just refer into the single-sample content of histfactory_dist as a way to encode data uncertainties with error sources, we propose a new type of distribution targeted to that usage. A complexity here is that nuisance parameters, constraints, etc. will have to be described in the data section, rather than nicely factorised into the models -- reality is messy!
  • UK and Poland reports mostly postponed to next time.
  • Turkey: EMCreator moving to Snakemake, some discussion on how to represent in-memory workflows that avoid writing big event files to disk (or doing the I/O at all); initial success with sumPCB surrogate modelling, need clarifications on which truth-level features are used and where the ATLAS-event ground-truth is from.
  • France: Arpita reported on exclusion differences between SModelS and Spey processing, to be investigated.
  •  
There are minutes attached to this event. Show them.