Abstract:We focus on a rare, exotic decay channel of the top quark, $t\to cH$, with $H$ decaying to a pair of b-quarks produced at the High Luminosity runs of the Large Hadron Collider. Our intent is to formulate a tagging algorithm which is efficient in identifying the said decay mode of the boosted top using a large-R jet with b- and c-tagged jets inside it. We consider the production of the top quark in association with a W-boson with the top decaying through $cH$ mode, and identify different observables as the effective discriminators of the signal over the Standard Model (SM) backgrounds, and finally perform a Multi Variate Analysis (MVA) using two different boosting algorithms, viz., XGBoost and AdaBoost. We show a comparative analysis on the performance of the proposed top tagger for the two boosting algorithms, and also with conventional cut-based top tagging algorithms. The algorithm we propose shows greater efficiency over the conventional ones. Furthermore, we make use of the mathematical formalism, SHAP, originated in the Game Theory to understand the role played by each observable towards the discriminative power of the classifier. This provides a great insight of the observables used to build the classifier. The tagging algorithm proposed in this work has wider range of applicability, it can be implemented in any probes of BSM Physics involving large-R, boosted jets with substructures involving b- and c-jets. As an illustration, we consider the Vector-like quark $Y$ decaying to b-quark and a $W$-boson which resembles the inner structure of the proposed toptagger. We observe clear resonance peak at the mass of the Y which proves that the proposed tagger is useful not only to tag the top quarks decaying through the above-mentioned flavour-violating mode, also is equally efficient in probing a BSM particle giving the same final state topology.

The Machine Learning Landscape of Top Taggers

Interplay of Traditional Methods and Machine Learning Algorithms for Tagging Boosted Objects

Top-philic Machine Learning

Accuracy versus precision in boosted top tagging with the ATLAS detector

Application of deep learning in top pair and single top quark production at the LHC

ML-Based Top Taggers: Performance, Uncertainty and Impact of Tower & Tracker Data Integration

Machine Learning in High Energy Physics: A review of heavy-flavor jet tagging at the LHC

From strange-quark tagging to fragmentation tagging with machine learning

Unveiling the Secrets of New Physics Through Top Quark Tagging

Application of Machine Learning Based Top Quark and W Jet Tagging to Hadronic Four-Top Final States Induced by SM and BSM Processes

Boosted top tagging and its interpretation using Shapley values

Application of Machine Learning Based Top Quark and W Jet Tagging to Hadronic Four-Top Final States Induced by SM as well as BSM Processes

Boosted Top Tagging through Flavour-violating interactions at the LHC

A Detailed Study of Interpretability of Deep Neural Network based Top Taggers

Jet Constituents for Deep Neural Network Based Top Quark Tagging

Exploring unsupervised top tagging using Bayesian inference

Pulling out all the tops with computer vision and deep learning

Novel machine learning applications at the LHC

Tau lepton identification and reconstruction: a new frontier for jet-tagging ML algorithms