Cargando…

Event Classification with Multi-step Machine Learning

<!--HTML-->The usefulness and valuableness of Multi-step ML, where a task is organized into connected sub-tasks with known intermediate inference goals, as opposed to a single large model learned end-to-end without intermediate sub-tasks, is presented. Pre-optimized ML models are connected and...

Descripción completa

Detalles Bibliográficos
Autor principal: Saito, Masahiko
Lenguaje:eng
Publicado: 2021
Materias:
Acceso en línea:http://cds.cern.ch/record/2766908
_version_ 1780971247608266752
author Saito, Masahiko
author_facet Saito, Masahiko
author_sort Saito, Masahiko
collection CERN
description <!--HTML-->The usefulness and valuableness of Multi-step ML, where a task is organized into connected sub-tasks with known intermediate inference goals, as opposed to a single large model learned end-to-end without intermediate sub-tasks, is presented. Pre-optimized ML models are connected and better performance is obtained by re-optimizing the connected one. The selection of a ML model from several small ML model candidates for each sub-task has been performed by using the idea based on NAS. In this paper, DARTS and SPOS-NAS are tested, where the construction of loss functions is improved to keep all ML models smoothly learning. Using DARTS and SPOS-NAS as an optimization and selection as well as the connecting for multi-step machine learning systems, we find that (1) such system can quickly and successfully select highly performant model combinations, and (2) the selected models are consistent with baseline algorithms such as grid search and their outputs are well controlled.
id cern-2766908
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2021
record_format invenio
spelling cern-27669082022-11-02T22:25:53Zhttp://cds.cern.ch/record/2766908engSaito, MasahikoEvent Classification with Multi-step Machine Learning25th International Conference on Computing in High Energy & Nuclear PhysicsConferences<!--HTML-->The usefulness and valuableness of Multi-step ML, where a task is organized into connected sub-tasks with known intermediate inference goals, as opposed to a single large model learned end-to-end without intermediate sub-tasks, is presented. Pre-optimized ML models are connected and better performance is obtained by re-optimizing the connected one. The selection of a ML model from several small ML model candidates for each sub-task has been performed by using the idea based on NAS. In this paper, DARTS and SPOS-NAS are tested, where the construction of loss functions is improved to keep all ML models smoothly learning. Using DARTS and SPOS-NAS as an optimization and selection as well as the connecting for multi-step machine learning systems, we find that (1) such system can quickly and successfully select highly performant model combinations, and (2) the selected models are consistent with baseline algorithms such as grid search and their outputs are well controlled.oai:cds.cern.ch:27669082021
spellingShingle Conferences
Saito, Masahiko
Event Classification with Multi-step Machine Learning
title Event Classification with Multi-step Machine Learning
title_full Event Classification with Multi-step Machine Learning
title_fullStr Event Classification with Multi-step Machine Learning
title_full_unstemmed Event Classification with Multi-step Machine Learning
title_short Event Classification with Multi-step Machine Learning
title_sort event classification with multi-step machine learning
topic Conferences
url http://cds.cern.ch/record/2766908
work_keys_str_mv AT saitomasahiko eventclassificationwithmultistepmachinelearning
AT saitomasahiko 25thinternationalconferenceoncomputinginhighenergynuclearphysics