Jet flavor tagging with Particle Transformer for Higgs factories

TL;DR

This paper evaluates the Particle Transformer model for jet flavor tagging in Higgs factory simulations, demonstrating significant improvements over previous methods and exploring potential for further enhancements with more data.

Contribution

The study introduces the application of Particle Transformer to jet flavor tagging, achieving substantial performance gains over traditional BDT-based methods.

Findings

5-10 times improvement in b/c tagging accuracy

Reasonable performance in strange tagging and quark/antiquark separation

Potential for further gains with increased training data

Abstract

We study the performance of the Particle Transformer (ParT) for jet flavor tagging using ILD full simulation events (1M jets) as well as fast simulation samples (10M and 1M jets). We perform 3-category ($b/c/d$), 6-category ($b/c/d/u/s/g$), and 11-category trainings (including quark--antiquark separation), incorporating multivariate hadron particle identification information from $dE/dx$ and time-of-flight. For $b$/$c$ tagging, we observe a factor of 5--10 improvement over previous BDT-based taggers, and we obtain reasonable performance for strange tagging and quark/antiquark separation. The 10M-jet fast simulation study indicates that further gains are possible with higher training statistics.