We provide different resources that help with the use of MadMiner:
Our main publication MadMiner: Machine-learning-based inference for particle physics provides an overview over this package. We recommend reading it first before jumping into the code.
In the examples folder in the MadMiner repository, we provide two tutorials. The first at examples/tutorial_toy_simulator/tutorial_toy_simulator.ipynb is based on a toy problem rather than a full particle-physics simulation. It demonstrates inference with MadMiner without spending much time on the more technical steps of running the simulation. The second, at examples/tutorial_particle_physics, shows all steps of a particle-physics analysis with MadMiner.
Typical work flow¶
Here we illustrate the structure of data analysis with MadMiner:
madminer.corecontains the functions to set up the process, parameter space, morphing, and to steer MadGraph and Pythia.
madminer.delphescontain two example implementations of a detector simulation and observable calculation. This part can easily be swapped out depending on the use case.
madminer.sampling, train and test samples for the machine learning part are generated and augmented with the joint score and joint ratio.
madminer.mlcontains an implementation of the machine learning part. The user can train and evaluate estimators for the likelihood ratio or score.
madminer.fisherinformationcontains functions to calculate the Fisher information, both on parton level or detector level, in the full process, individual observables, or the total cross section.
The madminer API is documented on here as well, just look through the pages linked on the left.