CaloGAN project

Please log in.

    • CaloGAN project

      In this challenge we ask you to train conditional generator of LHCb calorimeter response.

      The generator should model the following probability density:

      ρ(EnergyDeposit|ParticlePoint, ParticleMomentum, ParticlePDG).

       

      In order to submit your solution please follow the instructions below:

      1. Register in Coopetition, with the same email you have on github.

      2. Fork baseline repository, submit the link to your solution on github via MY SUBMISSIONS.

      3. If you want to see solutions of others checkout leaderboard of a competition to find a submission which might inspire you. Good luck!

       Also, I recommend to take a look at this presentation of one the participants of the competition: https://docs.google.com/presentation/d/1tvPfUBnC8mLV7I0P1AFzo_dIr6b61MDfLK0gVVC_q3c/edit?usp=sharing

       
    • Evaluation

      The submission will be evaluated using this code.

      Metric - Minimum of two PRD scores: 

      1. Over raw images (catches overal lproximity of distributions generated and real calorimeter responses)
      2. Over a set of physical metrics (catch esproximity of distributions for handmaded statistics)
      PRD score is a score that disentangles precision (quality of generated samples) from recall (proportion of target distribution that is covered by the generator, richness).
       
       
       
    • Rules

      Zipped results must be submitted before the 2020-05-01 20:09:00+00:00. You may submit 15 submissions every day and 1000 in total.

       
    •  

       
    • MLHEP 2019 stage #2: Data

      Dataset is in npz-format and could be opened with `np.load`. For example of how to work with it take a look on baseline.

       
    •  

       

      Development

      starting_kit public_data
  • Make your submission using github

    ID Status Inputs