Hypothetical supersymmetry particles linked to Higgs bosons (called Higgsinos) would create signatures in CMS where you have very low-momentum tracks plus unmeasurable particles like dark matter candidates. This #CMSPaper looks for that low-momentum+missing energy signature (with some mild machine learning). There is a non-significant deviation in this analysis, so we will look at it again with more data for sure! https://arxiv.org/abs/2511.16394
Post
Replies:
4
@freyablekman may I ask a practice question? How is the meta analysis with more data performed when the primary analysis has been conducted on various distributed compute environments in the first place?
I am asking, because we (some people from the #Snakemake community) recently had an online workshop with representatives of CERN. But I do not recall the meta analysis challenge being brought up. (But I might be mistaken.) And the Snakemake Hackathon is upcoming, where we want to invest some time into CERN matters.
@rupdecat as the tools are rapidly improving we tend to completely re-do the analysis on old data if there is some statistical gain. Otherwise the combination is done at statistical level (so combining the sensitive distributions of different years.). CMS has a dedicated statistical tool to make such non correlated distributions with correlated systematic uncertainties easier. Not creatively it’s called “combine” https://link.springer.com/article/10.1007/s41781-024-00121-4
@freyablekman thank you so much for this input! I will try to digest it later this week. 😊
@freyablekman PS I forgot to place a link to the meeting, I mentioned: https://indico.cern.ch/event/1643846/ . It took place online.
