Selective Combination and Management of Distributed Machine Learning Models

Takeshi Tsuchiya*, Ryuichi Mochizuki, Hiroo Hirose, Tetsuyasu Yamada, Keiichi Koyanagi, Quang Tran Minh

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This study presents a method for selecting and combining feature models constructed by the machine learning on the processing task capability. The evaluation of combining the feature models shows that the processing task capability can be improved by selecting and reaching feature models based on their similarity to the vector of queries without combining all feature models. Then, we discuss a method for constructing logical the R-Tree algorithm on the distributed fog nodes. For future work, we will implement the proposed method on various types of data.

Original languageEnglish
Title of host publicationFuture Data and Security Engineering - 8th International Conference, FDSE 2021, Proceedings
EditorsTran Khanh Dang, Josef Küng, Tai M. Chung, Makoto Takizawa
PublisherSpringer Science and Business Media Deutschland GmbH
Pages113-124
Number of pages12
ISBN (Print)9783030913861
DOIs
Publication statusPublished - 2021
Event8th International Conference on Future Data and Security Engineering , FDSE 2021 - Virtual, Online
Duration: 2021 Nov 242021 Nov 26

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13076 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference8th International Conference on Future Data and Security Engineering , FDSE 2021
CityVirtual, Online
Period21/11/2421/11/26

Keywords

  • Distributed future model
  • Fog computing model
  • Similarity of future models

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Selective Combination and Management of Distributed Machine Learning Models'. Together they form a unique fingerprint.

Cite this