Blekman Freya, Canelli Florencia, De Moor Alexandre, Gautam Kunal, Ilg Armin, Macchiolo Anna, Ploerer Eduardo
Inter-university Institute for High Energies, Vrije Universiteit Brussel, 1050 Brussels, Belgium.
Deutsches Elektronen-Synchrotron DESY, Notkestr. 85, 22607 Hamburg, Germany.
Eur Phys J C Part Fields. 2025;85(2):165. doi: 10.1140/epjc/s10052-025-13785-y. Epub 2025 Feb 10.
Jet flavour tagging is crucial in experimental high-energy physics. A tagging algorithm, DeepJet- Transformer, is presented, which exploits a transformer-based neural network that is substantially faster to train than state-of-the-art graph neural networks. The DeepJetTransformer algorithm uses information from particle flow-style objects and secondary vertex reconstruction for - and -jet identification, supplemented by additional information that is not always included in tagging algorithms at the LHC, such as reconstructed and and discrimination. The model is trained as a multiclassifier to identify all quark flavours separately and performs excellently in identifying - and -jets. An -tagging efficiency of can be achieved with a -jet background efficiency. The performance improvement achieved by including and reconstruction and discrimination is presented. The algorithm is applied on exclusive samples to examine the physics potential and is shown to isolate events. Assuming all non- backgrounds can be efficiently rejected, a discovery significance for can be achieved with an integrated luminosity of of collisions at , corresponding to less than a second of the FCC-ee run plan at the boson resonance.
喷注风味标记在实验高能物理中至关重要。本文提出了一种标记算法DeepJet-Transformer，它利用基于变换器的神经网络，其训练速度比当前最先进的图神经网络快得多。DeepJetTransformer算法使用来自粒子流风格对象和次级顶点重建的信息进行喷注识别，并辅以大型强子对撞机标记算法中并不总是包含的额外信息，例如重建的粲介子、底介子和轻子鉴别。该模型被训练为多分类器以分别识别所有夸克风味，并且在识别粲喷注和底喷注方面表现出色。可以在底喷注背景效率为的情况下实现的粲标记效率。展示了通过纳入粲介子和底介子重建以及轻子鉴别所实现的性能提升。该算法应用于专属样本以检验物理潜力，并被证明能够隔离粲事件。假设所有非粲背景都能被有效排除，在质心能量为的情况下，通过的碰撞积分亮度，对应于未来环形对撞机-电子正电子对撞机在玻色子共振处运行计划不到一秒的时间，可以实现的发现显著性。
