Speaker
Description
Jet tagging is a critical yet challenging classification task in particle physics. While deep learning has transformed jet tagging and significantly improved performance, the lack of a large-scale public dataset impedes further enhancement. In this work, we present JetClass, a new comprehensive dataset for jet tagging. The JetClass dataset consists of 100 M jets, about two orders of magnitude larger than existing public datasets. A total of 10 types of jets are simulated, including several types unexplored for tagging so far. Based on the large dataset, we propose a new Transformer-based architecture for jet tagging, called Particle Transformer (ParT). By incorporating pairwise particle interactions in the attention mechanism, ParT achieves higher tagging performance than a plain Transformer and surpasses the previous state-of-the-art, ParticleNet, by a large margin. The pre-trained ParT models, once fine-tuned, also substantially enhance the performance on two widely adopted jet tagging benchmarks.
https://arxiv.org/abs/2202.03772
Significance
Jet tagging is a widely adopted analysis technique in high energy physics experiment. In this work, we propose a large and comprehensive public dataset, JetClass. We also propose a transformer based machine learning model for jet tagging, Particle Transformer (ParT). Leveraging novel architect and special pairwise particle interaction information, ParT achieves state-of-the-art performance in jet tagging. Moreover, powered by the comprehensiveness and largeness, model pre-trained with JetClass performed better after fine-tuning with downstream tasks compared with the directly trained ones.
References
https://icml.cc/virtual/2022/poster/17989
https://indico.cern.ch/event/1144064/abstracts/144880/
https://indico.cern.ch/event/1078970/timetable/?view=standard#29-particle-transformer-for-je
