Cargando…

Real-Time Inference With 2D Convolutional Neural Networks on Field Programmable Gate Arrays for High-Rate Particle Imaging Detectors

We present a custom implementation of a 2D Convolutional Neural Network (CNN) as a viable application for real-time data selection in high-resolution and high-rate particle imaging detectors, making use of hardware acceleration in high-end Field Programmable Gate Arrays (FPGAs). To meet FPGA resourc...

Descripción completa

Detalles Bibliográficos
Autores principales: Jwa, Yeon-jae, Di Guglielmo, Giuseppe, Arnold, Lukas, Carloni, Luca, Karagiorgi, Georgia
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Frontiers Media S.A. 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9157595/
https://www.ncbi.nlm.nih.gov/pubmed/35664508
http://dx.doi.org/10.3389/frai.2022.855184
_version_ 1784718667719639040
author Jwa, Yeon-jae
Di Guglielmo, Giuseppe
Arnold, Lukas
Carloni, Luca
Karagiorgi, Georgia
author_facet Jwa, Yeon-jae
Di Guglielmo, Giuseppe
Arnold, Lukas
Carloni, Luca
Karagiorgi, Georgia
author_sort Jwa, Yeon-jae
collection PubMed
description We present a custom implementation of a 2D Convolutional Neural Network (CNN) as a viable application for real-time data selection in high-resolution and high-rate particle imaging detectors, making use of hardware acceleration in high-end Field Programmable Gate Arrays (FPGAs). To meet FPGA resource constraints, a two-layer CNN is optimized for accuracy and latency with KerasTuner, and network quantization is further used to minimize the computing resource utilization of the network. We use “High Level Synthesis for Machine Learning” (hls4ml) tools to test CNN deployment on a Xilinx UltraScale+ FPGA, which is an FPGA technology proposed for use in the front-end readout system of the future Deep Underground Neutrino Experiment (DUNE) particle detector. We evaluate network accuracy and estimate latency and hardware resource usage, and comment on the feasibility of applying CNNs for real-time data selection within the currently planned DUNE data acquisition system. This represents the first-ever exploration of employing 2D CNNs on FPGAs for DUNE.
format Online
Article
Text
id pubmed-9157595
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher Frontiers Media S.A.
record_format MEDLINE/PubMed
spelling pubmed-91575952022-06-02 Real-Time Inference With 2D Convolutional Neural Networks on Field Programmable Gate Arrays for High-Rate Particle Imaging Detectors Jwa, Yeon-jae Di Guglielmo, Giuseppe Arnold, Lukas Carloni, Luca Karagiorgi, Georgia Front Artif Intell Artificial Intelligence We present a custom implementation of a 2D Convolutional Neural Network (CNN) as a viable application for real-time data selection in high-resolution and high-rate particle imaging detectors, making use of hardware acceleration in high-end Field Programmable Gate Arrays (FPGAs). To meet FPGA resource constraints, a two-layer CNN is optimized for accuracy and latency with KerasTuner, and network quantization is further used to minimize the computing resource utilization of the network. We use “High Level Synthesis for Machine Learning” (hls4ml) tools to test CNN deployment on a Xilinx UltraScale+ FPGA, which is an FPGA technology proposed for use in the front-end readout system of the future Deep Underground Neutrino Experiment (DUNE) particle detector. We evaluate network accuracy and estimate latency and hardware resource usage, and comment on the feasibility of applying CNNs for real-time data selection within the currently planned DUNE data acquisition system. This represents the first-ever exploration of employing 2D CNNs on FPGAs for DUNE. Frontiers Media S.A. 2022-05-18 /pmc/articles/PMC9157595/ /pubmed/35664508 http://dx.doi.org/10.3389/frai.2022.855184 Text en Copyright © 2022 Jwa, Di Guglielmo, Arnold, Carloni and Karagiorgi. https://creativecommons.org/licenses/by/4.0/This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
spellingShingle Artificial Intelligence
Jwa, Yeon-jae
Di Guglielmo, Giuseppe
Arnold, Lukas
Carloni, Luca
Karagiorgi, Georgia
Real-Time Inference With 2D Convolutional Neural Networks on Field Programmable Gate Arrays for High-Rate Particle Imaging Detectors
title Real-Time Inference With 2D Convolutional Neural Networks on Field Programmable Gate Arrays for High-Rate Particle Imaging Detectors
title_full Real-Time Inference With 2D Convolutional Neural Networks on Field Programmable Gate Arrays for High-Rate Particle Imaging Detectors
title_fullStr Real-Time Inference With 2D Convolutional Neural Networks on Field Programmable Gate Arrays for High-Rate Particle Imaging Detectors
title_full_unstemmed Real-Time Inference With 2D Convolutional Neural Networks on Field Programmable Gate Arrays for High-Rate Particle Imaging Detectors
title_short Real-Time Inference With 2D Convolutional Neural Networks on Field Programmable Gate Arrays for High-Rate Particle Imaging Detectors
title_sort real-time inference with 2d convolutional neural networks on field programmable gate arrays for high-rate particle imaging detectors
topic Artificial Intelligence
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9157595/
https://www.ncbi.nlm.nih.gov/pubmed/35664508
http://dx.doi.org/10.3389/frai.2022.855184
work_keys_str_mv AT jwayeonjae realtimeinferencewith2dconvolutionalneuralnetworksonfieldprogrammablegatearraysforhighrateparticleimagingdetectors
AT diguglielmogiuseppe realtimeinferencewith2dconvolutionalneuralnetworksonfieldprogrammablegatearraysforhighrateparticleimagingdetectors
AT arnoldlukas realtimeinferencewith2dconvolutionalneuralnetworksonfieldprogrammablegatearraysforhighrateparticleimagingdetectors
AT carloniluca realtimeinferencewith2dconvolutionalneuralnetworksonfieldprogrammablegatearraysforhighrateparticleimagingdetectors
AT karagiorgigeorgia realtimeinferencewith2dconvolutionalneuralnetworksonfieldprogrammablegatearraysforhighrateparticleimagingdetectors