Speakers
Description
DUNE is a cutting edge experiment aiming to study neutrinos in detail, with a
special focus on the flavor oscillation mechanism. ProtoDUNE-SP (the prototype
of the DUNE Far detector Single Phase TPC), has been built and operated at CERN
and a full suite of reconstruction tools have been developed. Pandora is a
multi-algorithm framework that implements reconstructions tools: a large number of
algorithms (exploiting traditional clustering, detector physics and deep learning
approaches) are applied to images in order to gradually build up a picture of events.
The Pandora slicing algorithm aims to partition the detector hits of an
event in sets called slices. Each slice represents a single interaction in the
detector and should identify all the hits related to the interacting particle and its
subsequent decay products. We expect on the order of tens of slices per event in
ProtoDUNE-SP.
Here we present a deep learning approach to the problem, designing a model able to
outperform the state-of-the-art slicing algorithm which is currently implemented within
Pandora. We assess the performance of our tool in terms of efficiency and accuracy,
while exploiting hardware accelerating setups. The ultimate goal is to incorporate
this deep learning approach into the Pandora reconstruction.
Significance
The presentation covers novel results obtained in the reconstruction process, eventually the present work gives an idea of how far the new deep learning technologies can help developing software for the broad physics community
| Speaker time zone | Compatible with Europe |
|---|
