Jet tagging in the Lund plane with graph networks

Journal article

We study a bottom-up, holographic description of a field theory yielding the spontaneous breaking of an approximate SO(5) global symmetry to its SO(4) subgroup. The weakly-coupled, six-dimensional gravity dual has regular geometry. One of the dimensions is compactified on a circle that shrinks smoothly to zero size at a finite value of the holographic direction, hence introducing a physical scale in a way that mimics the effect of confinement in the dual four-dimensional field theory. We study the spectrum of small fluctuations of the bulk fields carrying SO(5) quantum numbers, which can be interpreted as spin-0 and spin-1 bound states in the dual field theory. This work supplements an earlier publication, focused only on the SO(5) singlet states. We explore the parameter space of the theory, paying particular attention to composite states that have the right quantum numbers to be identified as pseudo-Nambu-Goldstone Bosons (PNGBs). We find that in this model the PNGBs are generally heavy, with masses of the same order as other bound states, indicating the presence of a sizeable amount of explicit symmetry breaking in the field theory side. But we also find a qualitatively new, unexpected result. When the dimension of the field-theory operator inducing SO(5) breaking is close to half of the space-time dimensionality, there exists a region of parameter space in which the PNGBs and the lightest scalar are both parametrically light in comparison to all other bound states of the field theory. Although this region is known to yield metastable classical backgrounds, this finding might be relevant to model building in the composite Higgs context.Comment: 26 pages, 6 figures; Version accepted for publicatio

Authors: Dreyer, FA; Qu, H

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: Springer
• Journal: Journal of High Energy Physics
• Volume: 2021
• Issue: 3
• Pages: 52
• Article number: 52
• Publication date: 2021-03-04
• DOI: 10.1007/jhep03(2021)052
• EISSN: 1029-8479
• ISSN: 1126-6708
• Language: English
• Keywords: Artificial neural network; Overfitting; Geometry; Artificial intelligence; Jet (fluid); Computer science; Robustness (evolution); Theoretical computer science; Quark; Physics; Quantum mechanics; Large Hadron Collider; Mathematics; Scaling; Particle physics; Graph; Algorithm; Kinematics