Kun Liu weekly discussion with students [ATLAS+DarkSHINE]

Wednesday 9 Nov 2022, 16:00 → 17:35 Asia/Shanghai

Online

16:00 → 16:05 Introduction

General items.

Speaker: Prof. Kun LIU (Tsung-Dao Lee Institute & Shanghai Jiao Tong University)

16:05 → 16:15 ttWML and ttHML analyses status

Speaker: Nihal BRAHIMI (Tsung-Dao Lee Institute & Shanghai Jiao Tong University)

ttWML analysis status:

• Analysis was granted approval to unblind a month ago.
• However, a bug was found and fixed revealing a badly behaving inclusive fiducial fit compared to vanilla inclusive fit. This is only observed when moving to an improved sample (MG FxFx) for ttW Gen systematics compared to the previously used one, known to be suboptimal.
• Detailed tests were carried out but the cause of discrepancy between old and new MG FxFx sample has not been identified.
• Thorough discussion held at yesterday’s analysis meeting to decide on the steps forward for publication.
• The group decided to fall back to the sub-optimal MG FxFx sample for ttW Gen and target a conf-conversion for the LHCTop/HWG meeting in December: https://indico.cern.ch/event/1219500/
• After conf-conversion, we will proceed to a paper targeting Moriond.

ttHML analysis status:

• Most of the efforts focussed on defining MVA discriminants to be used in the 0-tau channels.
• Working with Jialin to come up with a good discriminant using DNN. Multiple configurations tested so far. 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. It seems that this level of correlation was already present in last round based on cross-checks from Chen (to be presented at today’s analysis meeting: https://indico.cern.ch/event/1220063/#3-mva-studies-updates-in-2lss).
• A first attempt at a stat-only combined fit including candidate SRs and fakes CRs from ttWML analysis was carried out. Level of ttW/ttH NFs correlation in 2lSS reduced but not significantly (down to 50%). More details can be found here: https://indico.cern.ch/event/1215174/contributions/5112688/attachments/2535731/4370621/NihalBRAHIMI_ttWMLMeeting_261022_bugFixed.pdf
• First dedicated (mini/skimmed) ttHML ntuples production rolled out end of last week and being carried out on the grid (Jialin volunteered for a subset of the samples).

16:15 → 16:30 ttH analysis

Speaker: jialin li

1109_kun.pdf

16:30 → 16:45 photon ID efficiency measurement & Zgamma analysis

Speaker: Kang Liu

For Rel21

I have finished the data17, mc16d, data16, mc16a grid jobs.

I'm running the condor job of Create finely-binned histograms for shower shape distributions for photons and electrons. (I submit it on 11/6, now more then 60% completed. this is the most time-consuming step).

After that I can verify whether I can get the same ID efficiency measurement results as before.

For Rel22

I added some required shower shape variables to the new TnP frame. Now I am trying to figure out the working point for electron and photon selection. 

For Silicon tracker

We have finished the IV test, and the CV test circuit has been set up. we will start the CV test this Friday.

 

16:45 → 17:00 silicon tracker: HGTD and DarkSHINE

paper draft with JINST template:

https://www.overleaf.com/4413756218tgbxvskzbcct

17:00 → 17:15 HCAL study

Speaker: Junhua Zhang

找到了truth动量的算法和变量位置

调研IV曲线的变量依赖

17:15 → 17:30 HCAL Study

Speaker: Hanqing Liu

To get a scale factor of incident Energy  /reco Energy of particle enter into HCAL

1.Finish the different particle(proton , neutron pi0, ki...) with energy range from(500-8000MeV) incident part.

17:30 → 17:35 Non-academic items

Anything else?