Jet Substructure Without Trees

TL;DR

To address jet substructure without relying on clustering histories, the paper introduces a clustering-free angular correlation function $\mathcal{G}(R)$ that encodes angular and mass scales from jet constituents. It then derives an angular structure function $\Delta \mathcal{G}(R)$ to identify prominent angular scales $R_*$ and associated partial masses $m_*$, yielding a suite of IRC-safe observables. As an application, it builds a top-quark tagger using these observables, achieving competitive performance against existing methods on BOOST2010 samples. The work suggests broad applicability to QCD studies and boosted-object tagging, and outlines future refinements to the observable set and peak identification approach.

Abstract

We present an alternative approach to identifying and characterizing jet substructure. An angular correlation function is introduced that can be used to extract angular and mass scales within a jet without reference to a clustering algorithm. This procedure gives rise to a number of useful jet observables. As an application, we construct a top quark tagging algorithm that is competitive with existing methods.

Figures (12)

• Figure 1: The angular correlation function $\mathcal{G}(R)$ for a sample top jet.
• Figure 2: $p_T$ plot and angular structure function $\Delta \mathcal{G}(R)$ for the top jet whose $\mathcal{G} (R)$ is illustrated in Fig. \ref{['gcurvetop']}. (a) The $p_T$ plot depicts the transverse energy deposited in calorimeter cells of size $0.1\times 0.1$ in $(\eta,\phi)$ with the area of each red square proportional to the $p_T$. This top has $p_T\sim 300$ GeV and a clean three-pronged substructure. (b) For a minimum prominence of 4.0, $\Delta \mathcal{G} (R)$ has three peaks with $R_{1*}=0.66$, $R_{2*}=0.91$, and $R_{3*}=1.48$. The red arrows illustrate the prominence of the two peaks at $R_{2*}$ and $R_{3*}$.
• Figure 3: An illustration of how prominence requirements, by selecting peaks that stand out above background noise, prevent angular scales from being double-counted.
• Figure 4: (a)$p_T$ plot and (b) angular structure function $\Delta \mathcal{G}(R)$ for a QCD jet with diffuse substructure and $p_T\sim600\ $GeV. In the $p_T$ plot, the small cell at the end of the arrow is so soft that it is barely visible. Prominent peaks in $\Delta \mathcal{G}(R)$ are distributed approximately uniformly in $R$. For a minimum prominence of 4.0, $\Delta \mathcal{G} (R)$ has a single peak at $R_{1*}=1.09$. Note the scale of $\Delta \mathcal{G}(R)$ as compared to the top jet in Fig. \ref{['masspeakplottop']}.
• Figure 5: (a)$p_T$ plot and (b) angular structure function $\Delta \mathcal{G}(R)$ for a top jet with $p_T \sim 500$ GeV. The decay products of the $W^\pm$ are not individually resolved, with most of the radiation from the $W^\pm$ ($\phi \sim 2.8$) contained within a single, hard cell. For a minimum prominence of 4.0, $\Delta \mathcal{G} (R)$ has a single peak at $R_{1*}=0.39$.