Speaker
Description
The use of Convolutional Neural Networks (CNN) techniques has grown
widely among the Liquid Argon Time Projection Chamber (LArTPC) community, mainly because the high-resolution images produced by these detectors are
suitable to be processed by such neural networks. Current and future LArTPC
experiments are constantly investigating different applications of CNNs as is the
case in the MicroBooNE and DUNE collaborations. In this poster, we present
preliminary results of using a CNN in the search of dark tridents in the Micro-
BooNE experiment. The dark trident signal is a new interaction channel that
allows a dark sector composed of a dark matter fermion and a dark photon that
could be explored in neutrino LArTPC experiments. We show that the CNN
achieves good discrimination of the signal from the background (NCπ 0 ), using
a simulated dataset. We include robustness checks, such as the performance
over different backgrounds, presence of cosmic ray activity in the dataset and
accuracy for different dark photon mass values.
