Reconstruction Group Meeting

Tuesday 27 Aug 2024, 15:30 → 16:40 US/Eastern

Zoom coordinates:

• URL: https://lbnl.zoom.us/j/94910574347?pwd=c01YQlg5bG04dFR6UWs3NjZkS3ZSdz09
• Meeting ID: 949 1057 4347
• Password: recon

ePIC reconstruction WG meeting
Aug 27, 2024

Host: Derek Anderson

Note taking: Shujie Li

Attendees: Sebouh, Rosi, Dmitri, Wouter

 

Derek proposed a new data structure to store track-cluster nad PID matching.

Discussions: 

Sebouh: there will be particle reconstructed without track e.g. in ZDC?
Wouter: you can only write the object once. 
Wouter: I do dislike the discrimination against neutrals in this model. Those poor track-less particles never get treated as a true (pseudo-)particle.
Rosi: need a structure for clusters of clusters, without tracking, for e.g. pi0
Wouter: The other part I don't like here is that you essentially are defining an omni-type that you want to have infinite flexibility. That's not what a data type is supposed to be.
Wouter: also missing the field for quality/likelyhood. 
Rosi: Maybe it makes sense to keep the pid a bit separate from the kinematics (i.e clusters/tracks)?
Sebouh: The name pseudoparticle already means something else. We should call it "ParticleCand"  https://www.quora.com/What-is-the-difference-between-quasiparticles-pseudoparticles-and-virtual-particles-or-do-they-all-mean-the-same-thing
Wouter: do something like mergingclusters instead of having a general data structure with field not filled. That makes downstream analysis too complicated.
Derek: on the PID side, it's an attempt to do calo-based PID which is not possible now.
Shujie: will we fill the track-ecal and track-hcal relation at once? 
Rosi: the real question is the one2many aspect. We may want to match track match to one cluster
Rosi: we can first do ecal-hcal matching, then track-cluster match
Derek: we need several stages: 1. track-ecal and track hcal, 2: ecal-hcal matching, 3: track-combined cluster
Derek: for track-cluster matching, how to convey the goodness of the algorithm.
Shujie: we can define a weighted distance between the track and the cluster center
Wouter: better use p value

15:30 → 15:40 Introduction¶

Speakers: Derek Anderson (Iowa State University), Shujie Li (Lawrence Berkeley National Laboratory)

Poll for New Meeting Time

PollForNewMeetingTime_epicRecoMeeting.d27m8y2024.v0.pdf

15:40 → 16:05 Track-Cluster Matching¶

Speakers: Derek Anderson (Iowa State University), Shujie Li (Lawrence Berkeley National Laboratory), Tristan Protzman (Lehigh University), Tyler Kutz (MIT)

FeedbackOnPseudoParticleProposal_epicRecoMeeting_28Aug2024.md

PseudoParticle Draft Proposal

16:05 → 16:20 [Postponed] Cross-Calorimeter Topocluster Update¶

Postponed to a later meeting

Speaker: Tristan Protzman (Lehigh University)

16:20 → 16:30 AOB¶