Separation of Left-Handed and Anomalous Right-Handed Vector Operators Contributions into the Wtb Vertex for Single and Double Resonant Top Quark Production Processes Using a Neural Network

The paper describes the application of deep neural networks for the search of deviations from the Standard Model predictions at the Wtb vertex in the processes of single and double resonant top quark production with identical final state tWb. Monte-Carlo events preliminary classified by first level neural network as corresponding to single or double resonant top quark production are analyzed by two second level neural networks if there is a possible contribution of the anomalous right-handed vector operator into Wtb vertex or events are corresponded to the Standard Model. The second level neural networks are different for single and double resonant classes. The classes depend differently on anomalous contribution and such splitting leads to better sensitivity. The developed statistical model is used to set constraints on the anomalous right-handed vector operator at the Wtb vertex in different regions of phase space. It is demonstrated that the proposed method allows to increase the efficiency of a search for the anomalous contributions to the Wtb vertex.