Simulated Di-Higgs samples used in 'Data-Driven High-Dimensional Statistical Inference with Generative Models'

Signal and background datasets for HI-SIGMA paper. The signal file contains ~200k events of standard model di-Higgs in the bb gamma gamma final state. The background is ~1M events of non-resonant production of the same final state (bb gamma gamma). Both sets of events have had a preselection applied described in the paper and then are stored in the h5 format. The 'bjets' and 'photons' fields contain the 4 vectors of the two b jets and two photons in the pt eta phi M and pt eta phi formats respectively. The 'feats' field contains the features used in the study and others. The entries in this field are: total_mass, Hbb_mass, Hgg_mass, Hbb_pt, Hgg_pt, dR_b1g1, dR_b1g2, dR_b2g1, dR_b2g2, dR_bb, dR_gg 'total_mass' is the total mass of the 4 object system. dR_b1g2 is the deltaR between the first b-jet and the 2nd photon. Madgraph cards used to generate these samples are also included. Please cite arxiv:XXXX