Skip to content
Training, generation, and analysis code for Learning Particle Physics by Example: Location-Aware Generative Adversarial Networks for Physics
Jupyter Notebook Python Makefile
Branch: master
Clone or download
Latest commit 88fee27 Aug 25, 2017
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
analysis add analysis folder with plotting notebook Feb 6, 2017
generation @ ab117cb adding generation sm Feb 7, 2017
models restructuring release Mar 19, 2017
.gitignore cleanup Nov 20, 2016
.gitmodules adding generation sm Feb 7, 2017
LICENSE update Lisence Nov 15, 2016
Makefile initial Nov 15, 2016
README.md Update README.md Aug 25, 2017

README.md

Location-Aware Generative Adversarial Networks (LAGAN) for Physics Synthesis

This repository contains all the code used in L. de Oliveira (@lukedeo), M. Paganini (@mickypaganini), B. Nachman (@bnachman), Learning Particle Physics by Example: Location-Aware Generative Adversarial Networks for Physics Synthesis [arXiv:1701.05927]

Citations

You are more than welcome to use the open data and open-source software provided here for any of your projects, but we kindly ask you that you please cite them using the DOIs provided below:

Asset Location
Source Code (this repository) DOI
Data (Pythia images) DOI
Model Weights DOI
Docker image (generation)

If you're using ideas or methods discussed in the paper, with or without using the software, please cite:

@article{lagan,
      author         = "de Oliveira, Luke and Paganini, Michela and Nachman, Benjamin",
      title          = "{Learning Particle Physics by Example: Location-Aware
                        Generative Adversarial Networks for Physics Synthesis}",
      year           = "2017",
      eprint         = "1701.05927",
      archivePrefix  = "arXiv",
      primaryClass   = "stat.ML",
      SLACcitation   = "%%CITATION = ARXIV:1701.05927;%%"
}

Getting Started

To clone everything necessary, you'll need to run git clone --recursive https://github.com/lukedeo/adversarial-jets to fetch all the submodules (you can add a -j6 or some other number to launch concurrent clones).

This repository is structured as such:

Generation

[TODO: allow for preprocessing in Docker] This folder links to the submodule used for generating Pythia images.

Models

This folder contains the Keras models used for training the LAGAN seen in the paper. By running python train.py -h from this folder, you should see all available options for running the training, as well as how to find / download the data required.

Analysis

[TODO: update filenames, download links, etc.] This folder contains a jupyter nootbook that will guide you through the production of the plots that appear in the paper. You will be able to reproduce them and modify them as you wish using our trained models and open datasets, or reuse the plotting functions to visualize the performance of your own LAGAN.

Simply run: jupyter notebook plots.ipynb

You can’t perform that action at this time.