Improved selective background Monte Carlo simulation at Belle II with graph attention networks and weighted events

TL;DR

This paper introduces an improved graph attention network-based filtering method for Monte Carlo simulations at Belle II, reducing computational costs by early event selection and addressing biases with statistical reweighting.

Contribution

It advances background filtering in Monte Carlo simulations using graph attention networks and develops statistical techniques to correct filtering biases.

Findings

Enhanced filter performance with graph attention networks

Effective bias correction through sampling and reweighting

Reduced computational resources for background simulation

Abstract

When measuring rare processes at Belle II, a huge luminosity is required, which means a large number of simulations are necessary to determine signal efficiencies and background contributions. However, this process demands high computation costs while most of the simulated data, in particular in case of background, are discarded by the event selection. Thus, filters using graph neural networks are introduced at an early stage to save the resources for the detector simulation and reconstruction of events discarded at analysis level. In our work, we improved the performance of the filters using graph attention and investigated statistical methods including sampling and reweighting to deal with the biases introduced by the filtering.