Suppression of accidental backgrounds with deep neural networks in the PandaX-II experiment

Abstract

Abstract The PandaX dark matter detection project searches for dark matter particles using the technology of dual phase xenon time projection chamber. The low expected rate of the signal events makes the control of backgrounds crucial for the experiment success. In addition to reducing external and internal backgrounds during the construction and operation of the detector, special techniques are employed to suppress the background events during the data analysis. In this article, we demonstrate the use of deep neural networks (DNNs) for suppressing the accidental backgrounds, as an alternative to the boosted-decision-tree method used in previous analysis of PandaX-II. A new data preparation approach is proposed to enhance the stability of the machine learning algorithms to be run and ultimately the sensitivity of the final data analysis.