BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:CIDeR-ML General Meeting
DTSTART:20260918T000000Z
DTEND:20260918T010000Z
DTSTAMP:20261009T004400Z
UID:indico-event-10734@indico.slac.stanford.edu
DESCRIPTION:https://u-tokyo-ac-jp.zoom.us/j/83932834349Link to recording\n
 \n\nQuick recap\nThe meeting focused on updates and discussions related to
  the development and application of SIREN models and other reconstruction 
 algorithms for WCTE data. Omar A. Alterkait reported on ongoing bug fixes 
 and rerunning of the reconstruction and calibration pipeline in LUCiD\, no
 ting some issues with photon contributions and edge cases. Zhenxiong prese
 nted preliminary results on beam reconstruction using WCT data\, comparing
  the performance of SIREN with other methods and outlining a draft structu
 re for a paper on SIREN applications in WCTE. Junjie shared detailed findi
 ngs on the limitations of SIREN models in learning high-frequency features
 \, particularly near PMT surfaces\, and discussed ablation studies on mode
 l architecture and omega zero settings. Yiyang briefly introduced a projec
 t on electron tracking using simulation-based inference\, highlighting pot
 ential improvements in angular resolution.\nNext steps\nJunjie\n\nFinish t
 he depth ablation for the SIREN model and determine the best combination o
 f omega_0 and depth for each target.\n\nOmar A. Alterkait\n\nFix bugs in L
 UCiD\, rerun the reconstruction and calibration pipeline\, and present res
 ults hopefully next week.\n\nPatrick\n\nFollow up on the MOU review status
  regarding the use of cosmic muon data for the SIREN paper.\n\nYiyang\n\nS
 end slides to Patrick or Kazuhiro and schedule a dedicated presentation fo
 r two weeks later.\n\nZhenxiong\n\nTry the pre-tuned SIREN model for beam 
 reconstruction and update the analysis accordingly.\nFinalize the draft of
  the SIREN paper for WCTE\, aiming for completion in one month\, and uploa
 d the outline link to the Indico page.\n\nSummary\nBeam Reconstruction Pro
 gress Update\nOmar A. Alterkait reported delays in presenting his work due
  to GPU issues but explained he is fixing bugs in LUCiD and rerunning the 
 reconstruction and calibration pipeline. The main bug identified affects p
 hoton contribution to sensors\, changing light amounts by 4%. Zhenxiong pr
 esented preliminary beam reconstruction results for WCTE using data from r
 un 1478\, comparing different reconstruction methods including OpticSIREN 
 and fiTQun. The results showed OpticSIREN achieved higher resolution than 
 fiTQun\, particularly in momentum reconstruction\, though some data select
 ion improvements may be needed.\nBimodal Structure and Data Reconstruction
 \nPatrick and Zhenxiong discussed the bimodal two-peak structure in vertic
 al direction plots\, with Zhenxiong noting uncertainty about whether this 
 was due to statistics and planning to explore it further. Ryo provided inf
 ormation about the current status of using fiTQun and LicketyFit algorithm
 s for data reconstruction\, indicating some people are testing these but t
 he implementation details are not fully established. Zhenxiong presented a
  draft outline for a SIREN paper for WCTE\, which will cover the demonstra
 tion of the SIREN model\, WCTE\, calibration and reconstruction using WCTE
  data\, including sections on the SIREN model\, WCTE detector\, cosmic muo
 n and beam selection\, and validation results.\nPaper Draft and MOU Review
 \nThe team discussed a paper draft that Zhenxiong estimated would take abo
 ut one month to complete. Junjie inquired about whether the paper would be
  a WCTE collaboration paper\, but Patrick clarified it would be a technica
 l paper with additional authorship. Patrick confirmed that the MOU review 
 deadline had passed two weeks ago with no comments\, indicating approval t
 o use the cosmic muon data.\nSIREN Prediction Performance Analysis\nJunjie
  presented findings on SIREN prediction performance limitations for a toy 
 detector\, analyzing frequency domain characteristics of different targets
  including visibility\, T0\, and PMT waveforms. The analysis revealed that
  PMT surfaces contribute significantly to high-frequency features\, with v
 isibility showing the strongest high-frequency content followed by PCA coe
 fficients and T0. Junjie conducted ablation studies comparing different mo
 del architectures and training approaches\, finding that separating T0 and
  visibility targets improved T0 prediction by 2-3 times\, and that larger 
 models with appropriate omega0 values are needed for PCA coefficient train
 ing due to higher frequency content.The team discussed how to select appro
 priate omega0 values for real data calibration\, with Junjie explaining th
 at omega0 is primarily determined by the type of data being processed rath
 er than specific calibration needs. Patrick and Junjie agreed that model d
 epth has a more significant impact on handling high-frequency features tha
 n omega0\, which controls convergence speed.\nYiyang briefly presented on 
 electron tracking with simulation-based inference\, discussing approaches 
 to improve angular resolution beyond marginalized general photo angle PDF 
 methods\, though the full presentation was deferred to a future meeting du
 e to time constraints.\n\n\n\nhttps://indico.slac.stanford.edu/event/10734
 /
URL:https://indico.slac.stanford.edu/event/10734/
END:VEVENT
END:VCALENDAR