A neural-network-defined Gaussian mixture model for particle identification applied to the LHCb fixed-target programme

Particle identification at high-energy physics experiments typically relies on classifiers combining different experimental observables. In this document, an innovative approach employing machine learning techniques to describe their dependence from the relevant features is presented. The proposed method is applied to the fixed-target programme at the LHCb experiment, where the sample size of the particle identification calibration channels affects the experimental performance. It is demonstrated to perform better than a model based on the LHCb detailed simulation and to be fast and suitable to a large variety of use cases.