We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

hep-ph

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

High Energy Physics - Phenomenology

Title: Convolutional Neural Networks for Direct Detection of Dark Matter

Authors: Charanjit K. Khosa (1), Lucy Mars (1), Joel Richards (1), Veronica Sanz (1, 2 and 3) ((1) Department of Physics and Astronomy, University of Sussex, Brighton BN1 9QH, UK, (2) Alan Turing Institute, British Library, 96 Euston Road, London NW1 2DB, UK and (3) Instituto de Física Corpuscular (IFIC), Universidad de Valencia-CSIC, E-46980, Valencia, Spain)
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\%.
Comments: 14 figures, 6 tables
Subjects: High Energy Physics - Phenomenology (hep-ph); Instrumentation and Methods for Astrophysics (astro-ph.IM)
Cite as: arXiv:1911.09210 [hep-ph]
  (or arXiv:1911.09210v1 [hep-ph] for this version)

Submission history

From: Charanjit Khosa [view email]
[v1] Wed, 20 Nov 2019 23:10:45 GMT (2491kb,D)

Link back to: arXiv, form interface, contact.