Attendees: Ashot, Hayian, Chao G., Chao P., Dipangkar, Mahbub, Weizhi, Xinzhan, Xuefei, Maxime, Mahbub

Weizhi:
- comparison of different target profiles
  -> negligible effects from uniform profile (considering profile close to COMSOL simulation)
- comparison of ep/ee super ratio between LG and PWO
  -> no big discrepancy found
- gem efficiency depending on logarithm parameter (not much on density correction)
- gem efficiency depending on gem pedestal cut -> 1% difference and angular dependency

Xinzhan:
- convolutional neural network implemented
- trained with data or simulation (fast convergence)
-> good results, can be merged with other NN algorithms

Xuefei:
- new noise adder in fitting software
- new fitter with Q6 and Q8 terms with fixed parameters
- larger uncertainties with Q6 and Q8 (effect below uncertainties)
- best fut with poly ratio independent from input
TASK: implement the continuous fraction fitter

Maxime:
- investigating energy wobbling
- looked at energy density profiles
  -> 1% raise at the edge of the modules