sPHENIX discussion of physics

[Murad Sarsour <murad.sarsour AT gmail.com>]

Hi Deepali,

Sorry! In their email it says 6/15 but it was a mistake - it was yesterday. Anyway, look up introductory slides which have lots of links and detailed information on how to proceed.

Regards,

Murad

On Jun 15, 2021, at 12:12 PM, Hideki Okawa <Hideki.Okawa AT cern.ch> wrote:

Heavy Flavor Topical Group Meeting (2021/6/15)

https://indico.bnl.gov/event/12007/ 

Introduction - Jin

Congratulations to Yasser, Cameron, Xuan and Sanghoon

for the RHIC/AGS Merit Awards and Zhaozhong for the 

poster award! 

Hideki: We'll need b/c-filtered jet samples as well, not just

   for the machine learning b-tagger, but also for the 

   substructure and D-jet correlation studies. 

Jin: Indeed. Will update the intro slides. 

HF tools - Cameron

p.5

Jin: Are these with all 50M events? 

Cameron: Yes. 

Jin: Statistics seems low. Better to double check. 

Zhaozhong: B+->D0(->K+ pi-) pi+ is a statistically challenging

   channel. B_s is even lower. By the way, are J/psi, psi(2s) from 

   the bbbar sample? 

Cameron: Yes. 

p.6

Wei: The asymmetry is not seen in other peaks. Is it due to the 

   eta cut? 

Cameron: No eta cut applied. B+ suffers from statistics. For J/psi,

   it could be due to the photon radiation. 

Wei: In principle, they should have the same asymmetry. 

Jin: Can we cut triple daughter events? 

Cameron: It is not straightforward.

p.7 

Jin: Better to double-check by applying |eta|<0.8, not just on 

   track 1, but also track 2. 

(Updates from Cameron sent to the mailing list after the meeting)

Xin: How is the precision of D0 mass position in the central region? 

Cameron: I was planning to have a look, but not yet. 

Jin: How was the performance in the previous workfest, where we had

   no mismatch in the magnetic field? 

Cameron: The simulation was simplified back then, so I am not sure 

  how reliable it would be to compare. 

Muon ID w/ machine learning - Ming

p.1 

Jin: Better to keep in mind the contributions from noise in the EMCal

   and HCal. 

p.11

Jin: Lots of hadrons from the background are not considered here yet. 

Ming: I need the MB sample. 

p.12

Hideki: What does "Early Stopping" mean here? 

Ming: It stops the training with a small set of train samples when the

   performance begins to saturate. 

Jin, Cameron, Zhaozhong: These functionalities (Random Forest, Boosted

Trees, etc.) exist in TMVA. 

Cameron: I will send you some code. 

Jin: Pytorch also has these functionalities implemented. 

Jin: Are 50~100M MB events sufficient? 

Ming: 100M should be decent. 

Cameron: We should also be careful with the over-training. 

Jin: For the electrons, it is better to also check the dE and dphi with the 

   3x3 clusters. 

Hideki: How is the cluster energy computed? Any cut on the noise? 

Jin: Basic noise and shape cuts are applied. 

Xin: What is the granularity of the HCal and EMCal? 

Jin: 0.1x0.1 for HCal, 0.025x0.025 for EMCal. 2x2 containes the shower 

in HCal and 3x3 for the EM shower. 

Xin: How is the occupancy? 

Jin: Ming's plots are with pp. HCal should be quite overwhelmed in AuAu. 

MDC1 vertexing performance - Dan 

p.3

Jin: Is this the number of reconstructed or truth tracks? 

Dan: Reco matched to the truth tracks from the PV.

Xin: Truth hits or reco hits? 

Dan: Reco hits. 

Xin: It would be better to make the x-axis in the log scale to see 

   the low pt part in detail. 

Hideki: Why are the errors significantly different between the x and z 

   resolution for Num. Tracks ~15-20? 

Dan: I need to check. The plots are made with TProfile. 

Jin: You should check the histograms in 1D. For the resolution, you can 

   use TH2::FitProfileY(). 

Hideki: Values in TProfile can visibly shift due to tails. You should check

   whether this is happening or not. 

Xin: What is the eta cut? 

Dan: |eta|<1.1

Jin: The vertex resolution looks bad. It seems that the tail effect is 

   sizable. 

Xin: What is the beam width? 

Jin: 100 microns

Xin: What about pileup? 

Jin: Not considered in these samples. 

Xin: We should consider the 2D decay length instead of 3D. The z 

   component may not work. 

Jin: Which sample are you using? 

Dan: Inclusive MB. I will also have a look at the HF samples. 

Jin: Do you apply a qhat cut? 

Dan: I don't know actually. Will check. 

Jin: Two checks to be done: (1) check the 1D histogram for # of tracks~5,

   (2) pick a few events and compare the truth and reco. There can be 

   mis-identification of the truth vertices. 

Jin: Where are the RAVE vertexing plots? 

Dan: I don't have the ones with MDC1 yet. 

Jin: What about the ones with your own Pythia samples? 

Dan: I have those plots and will add them to the slides after the meeting. 

   (Slides are now updated on the indico)

_______________________________________________
sPHENIX-physics-l mailing list
sPHENIX-physics-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/sphenix-physics-l