Login

Kun Liu weekly discussion with students [ATLAS]

→ Asia/Shanghai
Online

Online
    • 16:00 → 16:10
      ttWML and ttHML analyses status 10m
      Speaker: Nihal BRAHIMI (Tsung-Dao Lee Institute & Shanghai Jiao Tong University)

      ttWML analysis status:

      • Large sets of unblinded results (inclusive and differential cross-sections) already available and presented at the Top+X meeting: https://indico.cern.ch/event/1205638/#2-topq-2019-30-update-on-unbli
      • Overall positive feedback and requested checks are being followed-up
      • The team is now targeting to have the closure of the analysis at the beginning of January (after the holidays). This will grant us enough time to conclude on the remaining outstanding items and have the documentation (INT note and paper draft) ready.

      ttHML analysis status:

      • Efforts focussed on defining MVA discriminants to be used in the SRs
      • Working with Jialin to come up with a good discriminant using DNN. Multiple configurations tested. Main bottleneck is the ttW CR and ttH SR definitions. Large anti-correlation (70%) observed between NF(ttH) and NF(ttW) which is not desirable. We have decided to drop the 4D DNN option in favour of the 3D DNN setup from Chen (SDU). Studies performed will nevertheless be documented in the analysis INT note
      • First dedicated (mini/skimmed) ttHML ntuples production now mostly ready (few jobs to be retried). Main bottleneck for validating the samples conveniently is the limited eos space (HSG8 eos full). We are currently implementing a solution where the samples are saved on the Grid.
    • 16:10 → 16:25
      ttH analysis 15m
      Speaker: jialin li
      ttH_1214_kungroup.pdf
    • 16:25 → 16:40
      photon ID efficiency measurement & Zgamma analysis 15m
      Speaker: Kang Liu

      I have add a cut "pass__numerator__NOSYS == 1" for select probe electrons from Ntuples build by Rel22 TnP, and build a new Tree with shower shape variables. (For 16040 entries,  the number passed probe selection is 16008).

      next to do is that create finely-binned histograms for shower shape distributions, and calculate Smirnov transformations.

    • 16:40 → 17:00
      Zgamma combination 20m
      Speaker: yufeng wang

      Zy combination:

      • adding required classes into RooFitExtension
      • obtained single fit results (with no_multipdf ws)
      • trying ATLAS+CMS combination & studying the multipdf CMS ws
    • 17:00 → 17:15
      silicon tracker: HGTD 15m
    • 17:15 → 17:20
      Non-academic items 5m

      Anything else?