Studying the nature of the $P_c(4312)^+$ pentaquark state using Artificial Neural Networks

Pentaquark candidates have been seen by LHCb in the $J/\psi$p spectrum. One such candidate is the narrow $P_c(4312)^+$ state whose lineshape might be due to the opening of the $\Sigma^+_c\bar{D}^0$ threshold approximately 5 MeV higher. A recent analysis by the Joint Physics Analysis Center, JPAC, has suggested that the $P_c(4312)^+$ is likely to be a virtual state.

The amplitudes developed in the JPAC analysis are used to generate the mass spectrum that is used to train an artificial neural network, ANN, to classify the nature of the $P_c(4312)^+$ peak. Resolution and noise is also taken into consideration. ANNs are typically seen as black boxes but their inner workings can be understood by calculating the game theoretically optimal Shapley values. The details and performance of the ANN will be presented.