Punzi-loss, a non-differentiable metric approximation for sensitivity optimization in the search for new particles

Sumitted to PubDB: 2021-10-01

Category: Technical Paper, Visibility: Public

Tags: -

Authors	Paul Feichtinger, Gianluca Inguglia, James Kahn
Non-Belle II authors	H. Haigh
Date	2021-08-25
Belle II Number	BELLE2-PUB-TE-2021-001
Abstract	We present the novel implementation of a non-differentiable metric approximation and a corresponding loss-scheduling aimed at the search for new particles of unknown mass in high energy physics experiments. We call the loss-scheduling, based on the minimization of a figure-of-merit related function typical of particle physics, a Punzi-loss function, and the neural network that utilizes this loss function a Punzi-net. We show that the Punzi-net outperforms standard multivariate analysis techniques and generalizes well to mass hypotheses for which it was not trained. Our result constitutes a step towards fully differentiable analyses in particle physics. This work is implemented using PyTorch and we provide users full access to a public repository containing all the codes.