Speaker
Description
Identification of hadronic decays of highly Lorentz-boosted W/Z/Higgs bosons and top quarks provides powerful handles to a wide range of new physics searches and Standard Model measurements at the LHC. In this talk, we present ParticleNeXt, a new graph neural network (GNN) architecture tailored for jet tagging. With the introduction of novel components such as pairwise features, attentive pooling, and multi-scale aggregation in the GNN, the ParticleNeXt architecture achieves a significant performance improvement over state-of-the-art algorithms in several representative jet tagging tasks, including Higgs boson tagging, top quark tagging, and quark vs. gluon discrimination.
References
Presentation at ML4Jets2021: https://indico.cern.ch/event/980214/contributions/4413544/
Significance
The new algorithm presented in this talk, ParticleNeXt, outperforms state-of-the-art algorithms substantially in a broad range of jet tagging tasks.
| Speaker time zone | Compatible with Europe |
|---|
