Deep Learning for direct Dark Matter search with nuclear emulsions
We propose a new method for the discrimination of sub-micron nuclear recoil tracks from an instrumental background in fine-grain nuclear emulsions used in the directional dark matter search. The proposed method uses a 3D Convolutional Neural Network, whose parameters are optimised by Bayesian search. Unlike previous studies focused on extracting the directional information, we focus on the signal/background separation exploiting the polarisation dependence of the Localised Surface Plasmon Resonance phenomenon. Comparing the proposed method with the conventional cut-based approach shows a significant boost in the reduction factor for given signal efficiency.