Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userGraph Neural Networks for reconstruction and classification in KM3NeT
48
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
KM3NeT, a neutrino telescope currently under construction in the Mediterranean Sea, consists of a network of large-volume Cherenkov detectors. Its two different sites, ORCA and ARCA, are optimised for few GeV and TeV-PeV neutrino energies, respectively. This allows for studying a wide range of physics topics spanning from the determination of the neutrino mass hierarchy to the detection of neutrinos from astrophysical sources. Deep Learning techniques provide promising methods to analyse the signatures induced by charged particles traversing the detector. This document will cover a Deep Learning based approach using Graph Convolutional Networks to classify and reconstruct events in both the ORCA and ARCA detector. Performance studies on simulations as well as applications to real data will be presented, together with comparisons to classical approaches.
rate research
Read More
The KM3NeT research infrastructure is currently under construction at two locations in the Mediterranean Sea. The KM3NeT/ORCA water-Cherenkov neutrino detector off the French coast will instrument several megatons of seawater with photosensors. Its main objective is the determination of the neutrino mass ordering. This work aims at demonstrating the general applicability of deep convolutional neural networks to neutrino telescopes, using simulated datasets for the KM3NeT/ORCA detector as an example. To this end, the networks are employed to achieve reconstruction and classification tasks that constitute an alternative to the analysis pipeline presented for KM3NeT/ORCA in the KM3NeT Letter of Intent. They are used to infer event reconstruction estimates for the energy, the direction, and the interaction point of incident neutrinos. The spatial distribution of Cherenkov light generated by charged particles induced in neutrino interactions is classified as shower- or track-like, and the main background processes associated with the detection of atmospheric neutrinos are recognized. Performance comparisons to machine-learning classification and maximum-likelihood reconstruction algorithms previously developed for KM3NeT/ORCA are provided. It is shown that this application of deep convolutional neural networks to simulated datasets for a large-volume neutrino telescope yields competitive reconstruction results and performance improvements with respect to classical approaches.
The first prototype of a photo-detection unit of the future KM3NeT neutrino telescope has been deployed in the deep waters of the Mediterranean Sea. This digital optical module has a novel design with a very large photocathode area segmented by the use of 31 three inch photomultiplier tubes. It has been integrated in the ANTARES detector for in-situ testing and validation. This paper reports on the first months of data taking and rate measurements. The analysis results highlight the capabilities of the new module design in terms of background suppression and signal recognition. The directionality of the optical module enables the recognition of multiple Cherenkov photons from the same $^{40}$K decay and the localization bioluminescent activity in the neighbourhood. The single unit can cleanly identify atmospheric muons and provide sensitivity to the muon arrival directions.
Graph neural networks (GNNs) have received massive attention in the field of machine learning on graphs. Inspired by the success of neural networks, a line of research has been conducted to train GNNs to deal with various tasks, such as node classification, graph classification, and link prediction. In this work, our task of interest is graph classification. Several GNN models have been proposed and shown great accuracy in this task. However, the question is whether usual training methods fully realize the capacity of the GNN models. In this work, we propose a two-stage training framework based on triplet loss. In the first stage, GNN is trained to map each graph to a Euclidean-space vector so that graphs of the same class are close while those of different classes are mapped far apart. Once graphs are well-separated based on labels, a classifier is trained to distinguish between different classes. This method is generic in the sense that it is compatible with any GNN model. By adapting five GNN models to our method, we demonstrate the consistent improvement in accuracy and utilization of each GNNs allocated capacity over the original training method of each model up to 5.4% points in 12 datasets.
Directional detection of Dark Matter is a promising search strategy. However, to perform such detection, a given set of parameters has to be retrieved from the recoiling tracks : direction, sense and position in the detector volume. In order to optimize the track reconstruction and to fully exploit the data of forthcoming directional detectors, we present a likelihood method dedicated to 3D track reconstruction. This new analysis method is applied to the MIMAC detector. It requires a full simulation of track measurements in order to compare real tracks to simulated ones. We conclude that a good spatial resolution can be achieved, i.e. sub-mm in the anode plane and cm along the drift axis. This opens the possibility to perform a fiducialization of directional detectors. The angular resolution is shown to range between 20$^circ$ to 80$^circ$, depending on the recoil energy, which is however enough to achieve a high significance discovery of Dark Matter. On the contrary, we show that sense recognition capability of directional detectors depends strongly on the recoil energy and the drift distance, with small efficiency values (50%-70%). We suggest not to consider this information either for exclusion or discovery of Dark Matter for recoils below 100 keV and then to focus on axial directional data.
Muons created by $ u_mu$ charged current (CC) interactions in the water surrounding the ANTARES neutrino telescope have been almost exclusively used so far in searches for cosmic neutrino sources. Due to their long range, highly energetic muons inducing Cherenkov radiation in the water are reconstructed with dedicated algorithms that allow the determination of the parent neutrino direction with a median angular resolution of about unit{0.4}{degree} for an $E^{-2}$ neutrino spectrum. In this paper, an algorithm optimised for accurate reconstruction of energy and direction of shower events in the ANTARES detector is presented. Hadronic showers of electrically charged particles are produced by the disintegration of the nucleus both in CC and neutral current (NC) interactions of neutrinos in water. In addition, electromagnetic showers result from the CC interactions of electron neutrinos while the decay of a tau lepton produced in $ u_tau$ CC interactions will in most cases lead to either a hadronic or an electromagnetic shower. A shower can be approximated as a point source of photons. With the presented method, the shower position is reconstructed with a precision of about unit{1}{metre}, the neutrino direction is reconstructed with a median angular resolution between unit{2}{degree} and unit{3}{degree} in the energy range of SIrange{1}{1000}{TeV}. In this energy interval, the uncertainty on the reconstructed neutrino energy is about SIrange{5}{10}{%}. The increase in the detector sensitivity due to the use of additional information from shower events in the searches for a cosmic neutrino flux is also presented.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
