Bulletin of the American Physical Society
APS April Meeting 2022
Volume 67, Number 6
Saturday–Tuesday, April 9–12, 2022; New York
Session X09: Data Analysis, AI and ML IVRecordings Available
|
Hide Abstracts |
Sponsoring Units: DPF GDS Chair: Matthew Bellis, Siena College Room: Salon 3 |
Tuesday, April 12, 2022 10:45AM - 10:57AM |
X09.00001: MinGLE - Mini Geant4 Learning Example Jing Liu MinGLE, a Mini Geant4 Learning Example, uses minimal C++ coding (114 lines) to demonstrate the usage of essential Geant4 components step by step. It is not tied to any specific experiment or third party library, which makes it a clean starting point of writing your own Geant4 applications. |
Tuesday, April 12, 2022 10:57AM - 11:09AM |
X09.00002: Domain-informed neural networks for interaction localization within astroparticle experiments Shixiao Liang We propose a domain-informed neural network architecture for experimental particle physics, using particle interaction localization with the time-projection chamber (TPC) technology as an example application. While multilayer perceptrons (MLPs) have emerged as a leading contender for reconstruction in TPCs, this approach does not reflect prior knowledge of the underlying scientific processes. We encode prior detector knowledge, in terms of both signal characteristics and detector geometry, into the feature encoding and the output layers of a neural network. The resulting Domain-informed Neural Network (DiNN) limits the receptive fields of the neurons in the initial feature encoding layers to account for the spatially localized nature of the signals produced within the TPC, which significantly reduces the number of parameters in the network in comparison to an MLP. In addition, the output layers of the network are modified using two geometric transformations to ensure the DiNN produces localizations within the interior of the detector. The end result is a neural network architecture that has 60% fewer parameters than an MLP, but still achieves similar localization performance and provides a path to future architectural developments with improved performance. |
Tuesday, April 12, 2022 11:09AM - 11:21AM |
X09.00003: Explaining machine-learned particle-flow reconstruction Farouk Mokhtar, Raghav Kansal, Daniel C Diaz, Javier M Duarte, Joosep Pata, Maurizio Pierini, Jean-Roch Vlimant
|
Tuesday, April 12, 2022 11:21AM - 11:33AM |
X09.00004: ML-based Correction to Accelerate Geant4 Calorimeter Simulations Evangelos Kourlitis The Geant4 detector simulation, using full particle tracking (FullSim), is usually the most accurate detector simulation used in HEP but it is computationally expensive. The cost of FullSim is amplified in highly segmented calorimeters where large fraction of the computations are performed to track the shower's low-energy photons through the complex geometry. A method to limit the production of these photons is in the form of Geant4's production energy thresholds. Increasing the values of these thresholds reduces the accuracy of shower shapes in the simulation but can increase the computational speed. We propose a post-hoc machine learning (ML) correction method for calorimeter cell energy depositions. The method is based on learning the density ratio between the reduced accuracy simulation and the nominal one to extract multi-dimensional weights using a binary classifier. We explore the method using an example calorimeter geometry from the International Large Detector project and showcase initial results. The use of ML to correct calorimeter cells allows for more efficient use of heterogeneous computing resources with FullSim running on the CPU while the ML algorithm applies the correction in an event-parallel fashion on GPUs. |
Tuesday, April 12, 2022 11:33AM - 11:45AM |
X09.00005: Machine Learning Based Online Data Quality Monitoring for CMS ECAL Abhirami Harilal, Michael Andrews, Manfred Paulini The online Data Quality Monitoring (DQM) system of the CMS electromagnetic calorimeter (ECAL) is a vital operations tool. The DQM allows ECAL experts to quickly identify, localize, and diagnose a broad range of ECAL-related issues that would otherwise prevent the recording of physics-quality data. Continuous improvement has allowed the existing DQM system to be updated to respond to new problems; however, the aging of the ECAL electronics has resulted in rarer and more obscure failure modes, raising the need for a more robust anomaly detection system. Using unsupervised machine learning (ML), we have developed an auto encoder based anomaly detection system which can identify and localize anomalies within ECAL in real time. The auto encoder is robust with respect to changing detector conditions (eg, pile up) and takes into account the differential spatial responses and the time dependent nature of real anomalies. By periodically updating the spatial corrections and including time dependency, the efficiency of the DQM system is increased and false alarms are reduced. We will present this Run 3 production-quality ML-based online DQM system of the CMS ECAL. |
Tuesday, April 12, 2022 11:45AM - 11:57AM |
X09.00006: AutoDQM: A tool for monitoring data quality in the CMS detector John P Rotter, Chad Freer, Samuel May, Vivan Nguyen, Kaitlin Salyer, Si Sutantawibul, Robert White, Zhixing Che, Jonathan Guiang, Emanuele Barberis, Andrew Brinkerhoff, Indara Suarez, Darin Acosta AutoDQM is an automated monitoring system that utilizes statistical tests and machine learning (ML) algorithms to monitor the quality of data recorded by the CMS detector. It is used in conjunction with the existing Data Quality Monitoring (DQM) software to both reduce the time and labor required of shifters monitoring data quality and identify more subtle anomalies which may not be caught by eye. AutoDQM was used during the end of data-taking in Run 2 of the Large Hadron Collider (LHC) to monitor the Cathode Strip Chambers (CSC) and the Level-1 Trigger in the endcap region and has been expanded for Run 3 with the inclusion of ML algorithms including principal component analysis and deep autoencoders, which are being explored for their potential to improve the tool's ability to flag more subtle anomalies not flagged by statistical tests. During Run 3, the tool will accommodate more of the CMS detectors: Level-1 Trigger and several muon sub-detectors: CSC, Drift Tube chambers (DT), and Resistive Plate Chambers (RPC); however, is designed to be easily adaptable for use with other sub-detectors. High levels of modularity and a suite of tutorials allow experts from any CMS subsystem to leverage the functionalities of AutoDQM. |
Tuesday, April 12, 2022 11:57AM - 12:09PM |
X09.00007: Mu2e Event Display Development : Using the TEve and REve Frameworks Namitha Chithirasreemadam, Sophie Middleton, Simone Donati The Mu2e experiment will search for the CLFV neutrinoless coherent conversion of muon to electron, in the field of a nucleus. A custom Event Display has been developed using TEve, a ROOT based 3-D event visualisation framework (https://root.cern/doc/TEve). Event displays are crucial for monitoring and debugging during live data taking as well as for public outreach. A custom GUI allows event selection and navigation. Reconstructed data like the tracks, hits and clusters can be displayed within the detector geometries upon GUI request. True Monte Carlo trajectory of particles traversing the muon beam line, obtained directly from Geant4, can also be displayed. Tracks are coloured according to their particle identification and users get to select which trajectories to be displayed. Reconstructed tracks are refined using a Kalman filter. The resulting tracks can be displayed alongside truth information, allowing visualisation of the track resolution. The user can remove/add data based on energy deposited in a detector or arrival time. This is a prototype and an online event display, is currently under-development using the REve framework, which allows remote access for live data taking. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700