Mass Unspecific Supervised Tagging (MUST) for boosted jets

Abstract

Jet identification tools are crucial for new physics searches at the LHC and at future colliders. We introduce the concept of Mass Unspecific Supervised Tagging (MUST) which relies on considering both jet mass and transverse momentum varying over wide ranges as input variables — together with jet substructure observables — of a multivariate tool. This approach not only provides a single efficient tagger for arbitrary ranges of jet mass and transverse momentum, but also an optimal solution for the mass correlation problem inherent to current taggers. By training neural networks, we build MUST-inspired generic and multi-pronged jet taggers which, when tested with various new physics signals, clearly outperform the variables commonly used by experiments to discriminate signal from background. These taggers are also efficient to spot signals for which they have not been trained. Taggers can also be built to determine, with a high degree of confidence, the prongness of a jet, which would be of utmost importance in case a new physics signal is discovered.

Exotic diboson $$Z'$$ decays in the U$$\mu \nu $$SSM

“…Because we are not looking for a specific signal, a narrow jet mass window cannot be a priori imposed. Neither a dedicated tagger can be used; rather, a generic tagger for multi-pronged jets [27][28][29], or generic anomaly detection methods [30][31][32][33][34][35][36][37][38][39][40][41][42] must be used. In order to take advantage of the presence of b quarks, the sample is divided into different categories corresponding to the number of b tags, and the results are subsequently combined.…”

“…We use the MUST inspired tagger GenT in Ref. [29] that classifies any type of multi-pronged jets as signal, and QCD jets as background. The NN score X is shown in Fig.…”

Better latent spaces for better autoencoders

“…Additional papers studying similar unsupervised LHC problems include Ref. [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40].…”

Enhancing Jet Taggers with Mass Unspecific Supervised Tagging (MUST)

“…We adopt a different strategy in this work, in which m J and p T become input variables varying over wide ranges. This new approach that we designate by Mass Unspecific Supervised Tagging (MUST) [33] provides a great performance across wide ranges of m J and p T , solving simultaneously the mass correlation problem in the best possible way by preserving the shape of the m J distribution. Moreover, like other supervised taggers [31,32], those based on MUST can be generic, in the sense that they can discriminate multiple types of signal from background.…”

“…In this paper we just present the most important properties of the datasets we use to train our taggers. For more details about their generation, refer to[33].…”