Skip to content

LukeV37/Pileup_Project

Repository files navigation

Quick Start

As a user, clone the repo over https using:

git clone --recursive https://github.com/LukeV37/Pileup_Project.git

As a developer, clone the repo over ssh using:

git clone --recursive git@github.com:LukeV37/Pileup_Project.git

Ensure dependencies are installed (see below), and build the submodules:

./build_submodules.sh

Please be patient while submodules build...

See software/README.md for more information regarding submodules.

How To Generate Datasets

First, generate events in MadGraph:

cd madgraph
# Generate Signal
./run_dihiggs.sh
# Generate Background
./run_4b.sh

For more information on pythia, please see madgraph/README.md.

Then, shower in Pythia8:

cd pythia
./generate_diHiggs.sh {diHiggs|4b} {mu} {String Modifier} {MinJetpT}

The root file generated by the pythia simulation will be found in pythia/output.

For more information on pythia, please see pythia/README.md.

Folder structure

  • madgraph/: Generate Events in LHE format using MadGraph5
    • run_dihiggs.sh: Generate signal process
    • run_4b.sh: Generate background process
  • pythia/: Shower Events to ROOT files using Pythia8
    • generate_diHiggs.sh
  • models/: Train a model to predict Efrac and Mfrac of jets
    • run_preprocessing.sh: Convert ROOT file to Pytorch Tensors and perform basic cuts
    • train_model.sh: Train the model on the preprocessed data to predict Efrac and Mfrac
  • analysis/: Distinguish between diHiggs and 4b using Efrac and Mfrac
    • run_preprocessing.sh: Convert ROOT file to Pytorch Tensors and perform basic cuts
    • train_model.sh: Train the model on the preprocessed data to distinguish between diHiggs and 4b
  • software/: Git references to external packages
    • install_scripts/: Folder containing scripts to auto-install

Dependencies

Tested on Ubuntu 22.04, EL9 lxplus machine, and even WSL2.

Required Dependencies:

  • python3
  • ROOTv6
  • autoconf
  • libtool
  • automake
  • tcl
  • gfortran
  • g++

ROOTv6 can be installed following instructions here.

Other packages can be intalled with sudo apt install {package}

Misc.

Warning

ROOT must be installed on the system before ./install.sh script can be run.
Use root-config --cflags --libs to see if you have a successful ROOT install.

Note

On GPURIG2, please add the following lines to your .bashrc file in your home directory:

export ROOTSYS=/usr/local/rootpy27/root_install
export PATH=$ROOTSYS/bin:$PATH

About

A PhD project using Attention and Transformer Encoder stacks for a regression task to predict the energy and mass contributions from pileup on per jet basis in the context of an event. Generated in MadGraph, showered in Pythia8, clustered using FastJet3.

Resources

Stars

0 stars

Watchers

1 watching

Forks

Contributors