Cargando…

Hyperparameter optimization, quantum-assisted model performance prediction, and benchmarking of AI-based High Energy Physics workloads using HPC

Training and Hyperparameter Optimization (HPO) of deep learning-based AI models are often compute resource intensive and calls for the use of large-scale distributed resources as well as scalable and resource efficient hyperparameter search algorithms. This work studies the potential of using model...

Descripción completa

Detalles Bibliográficos
Autores principales: Wulff, Eric, Girone, Maria, Southwick, David, Amboage, Juan Pablo García, Cuba, Eduard
Lenguaje:eng
Publicado: 2023
Materias:
Acceso en línea:http://cds.cern.ch/record/2859790
_version_ 1780977710830452736
author Wulff, Eric
Girone, Maria
Southwick, David
Amboage, Juan Pablo García
Cuba, Eduard
author_facet Wulff, Eric
Girone, Maria
Southwick, David
Amboage, Juan Pablo García
Cuba, Eduard
author_sort Wulff, Eric
collection CERN
description Training and Hyperparameter Optimization (HPO) of deep learning-based AI models are often compute resource intensive and calls for the use of large-scale distributed resources as well as scalable and resource efficient hyperparameter search algorithms. This work studies the potential of using model performance prediction to aid the HPO process carried out on High Performance Computing systems. In addition, a quantum annealer is used to train the performance predictor and a method is proposed to overcome some of the problems derived from the current limitations in quantum systems as well as to increase the stability of solutions. This allows for achieving results on a quantum machine comparable to those obtained on a classical machine, showing how quantum computers could be integrated within classical machine learning tuning pipelines. Furthermore, results are presented from the development of a containerized benchmark based on an AI-model for collision event reconstruction that allows us to compare and assess the suitability of different hardware accelerators for training deep neural networks.
id cern-2859790
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2023
record_format invenio
spelling cern-28597902023-06-07T02:37:57Zhttp://cds.cern.ch/record/2859790engWulff, EricGirone, MariaSouthwick, DavidAmboage, Juan Pablo GarcíaCuba, EduardHyperparameter optimization, quantum-assisted model performance prediction, and benchmarking of AI-based High Energy Physics workloads using HPCcs.LGComputing and Computersphysics.data-anOther Fields of PhysicsTraining and Hyperparameter Optimization (HPO) of deep learning-based AI models are often compute resource intensive and calls for the use of large-scale distributed resources as well as scalable and resource efficient hyperparameter search algorithms. This work studies the potential of using model performance prediction to aid the HPO process carried out on High Performance Computing systems. In addition, a quantum annealer is used to train the performance predictor and a method is proposed to overcome some of the problems derived from the current limitations in quantum systems as well as to increase the stability of solutions. This allows for achieving results on a quantum machine comparable to those obtained on a classical machine, showing how quantum computers could be integrated within classical machine learning tuning pipelines. Furthermore, results are presented from the development of a containerized benchmark based on an AI-model for collision event reconstruction that allows us to compare and assess the suitability of different hardware accelerators for training deep neural networks.arXiv:2303.15053oai:cds.cern.ch:28597902023-03-27
spellingShingle cs.LG
Computing and Computers
physics.data-an
Other Fields of Physics
Wulff, Eric
Girone, Maria
Southwick, David
Amboage, Juan Pablo García
Cuba, Eduard
Hyperparameter optimization, quantum-assisted model performance prediction, and benchmarking of AI-based High Energy Physics workloads using HPC
title Hyperparameter optimization, quantum-assisted model performance prediction, and benchmarking of AI-based High Energy Physics workloads using HPC
title_full Hyperparameter optimization, quantum-assisted model performance prediction, and benchmarking of AI-based High Energy Physics workloads using HPC
title_fullStr Hyperparameter optimization, quantum-assisted model performance prediction, and benchmarking of AI-based High Energy Physics workloads using HPC
title_full_unstemmed Hyperparameter optimization, quantum-assisted model performance prediction, and benchmarking of AI-based High Energy Physics workloads using HPC
title_short Hyperparameter optimization, quantum-assisted model performance prediction, and benchmarking of AI-based High Energy Physics workloads using HPC
title_sort hyperparameter optimization, quantum-assisted model performance prediction, and benchmarking of ai-based high energy physics workloads using hpc
topic cs.LG
Computing and Computers
physics.data-an
Other Fields of Physics
url http://cds.cern.ch/record/2859790
work_keys_str_mv AT wulfferic hyperparameteroptimizationquantumassistedmodelperformancepredictionandbenchmarkingofaibasedhighenergyphysicsworkloadsusinghpc
AT gironemaria hyperparameteroptimizationquantumassistedmodelperformancepredictionandbenchmarkingofaibasedhighenergyphysicsworkloadsusinghpc
AT southwickdavid hyperparameteroptimizationquantumassistedmodelperformancepredictionandbenchmarkingofaibasedhighenergyphysicsworkloadsusinghpc
AT amboagejuanpablogarcia hyperparameteroptimizationquantumassistedmodelperformancepredictionandbenchmarkingofaibasedhighenergyphysicsworkloadsusinghpc
AT cubaeduard hyperparameteroptimizationquantumassistedmodelperformancepredictionandbenchmarkingofaibasedhighenergyphysicsworkloadsusinghpc