Convolutional neural networks for direct detection of dark matter

Abstract

The XENON1T experiment uses a time projection chamber (TPC) with liquid Xenon to search for Weakly Interacting Massive Particles (WIMPs), a proposed Dark Matter particle, via direct detection. As this experiment relies on capturing rare events, the focus is on achieving a high recall of WIMP events. Hence the ability to distinguish between WIMP and the background is extremely important. To accomplish this, we suggest using Convolutional Neural Networks (CNNs); a Machine Learning procedure mainly used in image recognition tasks. To explore this technique we use XENON collaboration open-source software to simulate the TPC graphical output of Dark Matter signals and main backgrounds. A CNN turns out to be a suitable tool for this purpose, as it can identify features in the images that differentiate the two types of events without the need to manipulate or remove data in order to focus on a particular region of the detector. We find that the CNN can distinguish between the dominant background events (ER) and 500 GeV WIMP events with an accuracy of 87.0% and a recall of 88.2%.

Tagging the Higgs boson decay to bottom quarks with colour-sensitive observables and the Lund jet plane

“…CNNs are showing promising potential for image based data sets for various applications in HEP see e.g. [31][32][33].…”

A semi-supervised approach to dark matter searches in direct detection data with machine learning

“…We note, however, that on an event-by-event basis, the fundamental detector response to a NR vs ER is different, and therefore supervised classification techniques can be used to discriminate between signal and background, as was observed in Ref. [43]. A key result we found whilst validating the results of this study was that such supervised methods work exceptionally well for background rejection regardless of the DM mass and cross-section as well, since the fundamental property being learned by the model is the NR and ER detector response.…”

“…Interestingly, in Ref. [34] a convolution neural network was applied to XENON1T TPC detector response images of WIMP signal events (nuclear recoils) and background events (electron recoils) achieving a classification accuracy of above 90%. In this study, we exploit a pre-trained convolutional neural network with a similar architecture to boost the anomaly awareness of our unsupervised model.…”

Event vertex reconstruction with deep neural networks for the DarkSide-20k experiment

“…Modern deep learning methods are being increasingly applied for event reconstruction in neutrino experiments such as MicroBooNE [3,4] and DUNE [5,6]. Moreover, studies based on convolution neural networks (CNNs) are also found for dual-phase TPC-based experiments, for example, the EXO-200 [7], the XENONnNT [8], and the DarkSide-50 [9] collaborations. In addition to these, several applications of various neural network techniques were demonstrated for the LUX xenon TPC [10], not covering direct XY vertex reconstruction.…”