اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدNovel deep learning methods for track reconstruction
280
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
For the past year, the HEP.TrkX project has been investigating machine learning solutions to LHC particle track reconstruction problems. A variety of models were studied that drew inspiration from computer vision applications and operated on an image-like representation of tracking detector data. While these approaches have shown some promise, image-based methods face challenges in scaling up to realistic HL-LHC data due to high dimensionality and sparsity. In contrast, models that can operate on the spacepoint representation of track measurements (hits) can exploit the structure of the data to solve tasks efficiently. In this paper we will show two sets of new deep learning models for reconstructing tracks using space-point data arranged as sequences or connected graphs. In the first set of models, Recurrent Neural Networks (RNNs) are used to extrapolate, build, and evaluate track candidates akin to Kalman Filter algorithms. Such models can express their own uncertainty when trained with an appropriate likelihood loss function. The second set of models use Graph Neural Networks (GNNs) for the tasks of hit classification and segment classification. These models read a graph of connected hits and compute features on the nodes and edges. They adaptively learn which hit connections are important and which are spurious. The models are scaleable with simple architecture and relatively few parameters. Results for all models will be presented on ACTS generic detector simulated data.
قيم البحث
اقرأ أيضاً
Muons are the most abundant charged particles arriving at sea level originating from the decay of secondary charged pions and kaons. These secondary particles are created when high-energy cosmic rays hit the atmosphere interacting with air nuclei ini
tiating cascades of secondary particles which led to the formation of extensive air showers (EAS). They carry essential information about the extra-terrestrial events and are characterized by large flux and varying angular distribution. To account for open questions and the origin of cosmic rays, one needs to study various components of cosmic rays with energy and arriving direction. Because of the close relation between muon and neutrino production, it is the most important particle to keep track of. We propose a novel tracking algorithm based on the Geometric Deep Learning approach using graphical structure to incorporate domain knowledge to track cosmic ray muons in our 3-D scintillator detector. The detector is modeled using the GEANT4 simulation package and EAS is simulated using CORSIKA (COsmic Ray SImulations for KAscade) with a focus on muons originating from EAS. We shed some light on the performance, robustness towards noise and double hits, limitations, and application of the proposed algorithm in tracking applications with the possibility to generalize to other detectors for astrophysical and collider experiments.
This paper presents a novel neutral-pion reconstruction that takes advantage of the machine learning technique of semantic segmentation using MINERvA data collected between 2013-2017, with an average neutrino energy of $6$ GeV. Semantic segmentation
improves the purity of neutral pion reconstruction from two gammas from 71% to 89% and improves the efficiency of the reconstruction by approximately 40%. We demonstrate our method in a charged current neutral pion production analysis where a single neutral pion is reconstructed. This technique is applicable to modern tracking calorimeters, such as the new generation of liquid-argon time projection chambers, exposed to neutrino beams with $langle E_ u rangle$ between 1-10 GeV. In such experiments it can facilitate the identification of ionization hits which are associated with electromagnetic showers, thereby enabling improved reconstruction of charged-current $ u_e$ events arising from $ u_{mu} rightarrow u_{e}$ appearance.
Recently, deep learning approaches have become the main research frontier for biological image reconstruction problems thanks to their high performance, along with their ultra-fast reconstruction times. However, due to the difficulty of obtaining mat
ched reference data for supervised learning, there has been increasing interest in unsupervised learning approaches that do not need paired reference data. In particular, self-supervised learning and generative models have been successfully used for various biological imaging applications. In this paper, we overview these approaches from a coherent perspective in the context of classical inverse problems, and discuss their applications to biological imaging.
Machine learning has proven to be an indispensable tool in the selection of interesting events in high energy physics. Such technologies will become increasingly important as detector upgrades are introduced and data rates increase by orders of magni
tude. We propose a toolkit to enable the creation of a drone classifier from any machine learning classifier, such that different classifiers may be standardised into a single form and executed in parallel. We demonstrate the capability of the drone neural network to learn the required properties of the input neural network without the use of any labels from the training data, only using appropriate questioning of the input neural network.
We explore the use of Deep Learning to infer physical quantities from the observable transmitted flux in the Lyman-alpha forest. We train a Neural Network using redshift z=3 outputs from cosmological hydrodynamic simulations and mock datasets constru
cted from them. We evaluate how well the trained network is able to reconstruct the optical depth for Lyman-alpha forest absorption from noisy and often saturated transmitted flux data. The Neural Network outperforms an alternative reconstruction method involving log inversion and spline interpolation by approximately a factor of 2 in the optical depth root mean square error. We find no significant dependence in the improvement on input data signal to noise, although the gain is greatest in high optical depth regions. The Lyman-alpha forest optical depth studied here serves as a simple, one dimensional, example but the use of Deep Learning and simulations to approach the inverse problem in cosmology could be extended to other physical quantities and higher dimensional data.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
