A Appendix 451 A.1 Shower shape variables

We extend the list of shower shape variables described in Sec. Figure 1 shows the average events for different variations of SUP A datasets and Figs. Figure 1 - 12 shows PointFlow PointFlow [Y ang et al., 2019] is a flow based model with a PointNet-like encoder and a The overall architecture has 2 . 1 M parameters. We train all the models with 100 K training examples. Figs. 13 - 18 show the histograms of various shower shape variables for SUP Av1 and samples Figure 18: Histograms of various shower shape variables A.4.2 SUPA v2 Figs. We can increase the number of cells in order to get higher resolutions. We choose 1x to be the same resolution as used in CaloGAN [Paganini et al., In this section, we extend those studies with more analysis and plots.