Deep learning for $R$-parity violating supersymmetry searches at the LHC

Supersymmetry with hadronic R-parity violation is still weakly constrained, in which the lightest neutralino can decay into three quarks. We construct a Convolutional Neutral Network (CNN) which is capable to tag the neutralino jet in any signal processes by using the idea of jet image. Such CNN outperforms the jet substructure variable N-subjettiness by a factor of a few in tagging efficiency. Combining with the jet invariant mass, our CNN can perform better and is applicable to a wider range of neutralino mass. The ATLAS search for the signal of gluino pair production with subsequent decay $\tilde{g} \to q q \tilde{\chi}^0_1 (\to q q q)$ is recasted as a detailed application. Supplied with the additional information from our CNN, the exclusion limit on gluino mass can be pushed up by $\sim200$ GeV.