Calorimeter Machine Learning

Friday 1 Jun 2018, 15:00 → 18:00 Europe/Zurich

target publication for end of summer at the latest

to do:

not enough chPi events, need more - Maurizio is generating

replicate NIPS results with new samples (fixed and variable separately) with original window size and architecture

complete and write up studies on impact of window size on classification accuracy

Dominick found some variable-angle samples had very low energy in ECAL - visualize these events and see why

look at Kyle's paper (https://arxiv.org/pdf/1601.07913.pdf) and get more ideas for how to feed eta into classifier

check modified GAN architecture for inputting eta and seeing if improvements can be made

15:00 → 15:20 Maurizio

Speaker: Maurizio Pierini (CERN)

15:20 → 15:40 Ben

Speaker: Benjamin Henry Hooberman (Univ. Illinois at Urbana-Champaign (US))

15:40 → 16:00 Amir

Speaker: Amir Farbin (University of Texas at Arlington (US))

16:00 → 16:20 Jean-Roch

Speaker: Dr Jean-Roch Vlimant (California Institute of Technology (US))

16:20 → 16:40 Ryan

Speaker: Ryan Reece (University of California,Santa Cruz (US))

16:40 → 17:00 Matt

Speaker: Matt Zhang (Univ. Illinois at Urbana-Champaign (US))

17:00 → 17:20 Taylor

17:20 → 17:40 Junze

17:40 → 18:00 Dominick - Regression

Speaker: Dominick Olivito (Univ. of California San Diego (US))

lcdml_reg_olivito_010618.pdf