Falcon is an ultra-fast non-parametric detector simulation package that automatically abstracts detector response, usually done by hand in fast-simulators used by particle physics experiments. Falcon uses KDTrees to build a fast lookup table to map events at the parton shower level to events at the reconstruction level as described in the following paper.
The goal of this project is to optimize the structure of the code by extending the KDTreeBinning class to include automatic re-partitioning as more points are added (ie. refining each bin) and capability to rapidly and efficiently access points inside of each partition. Additionally, the goal is to integrate multi-target regression capability into Falcon.
Strong development skills, good knowledge of C++ and Python. Interest in Machine Learning algorithms and applications.