• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

A Study on the Configuration of Autonomous Cooperative Edge AI Systems

Research Project

Project/Area Number 20K11734
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 60040:Computer system-related
Research InstitutionKyoto Institute of Technology

Principal Investigator

NUNOME Atsushi  京都工芸繊維大学, 情報工学・人間科学系, 准教授 (60335320)

Co-Investigator(Kenkyū-buntansha) 平田 博章  京都工芸繊維大学, 情報工学・人間科学系, 教授 (90273549)
Project Period (FY) 2020-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2022: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2021: ¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000)
Fiscal Year 2020: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Keywords計算機システム / エッジコンピューティング / 分散協調処理 / 高性能計算
Outline of Research at the Start

現在のエッジAI処理では、低遅延で学習推論成果を得られる反面、クラウドサーバ上で行う学習処理よりも精度が劣ってしまうという課題があった。今後、エッジデバイスのハードウェアが進化するにしたがって、エッジデバイス上での学習精度はある程度向上することが予想されるものの、エッジAIがもつ課題を本質的に解決するとは言いがたい。本研究では、複数のエッジデバイスが自律的に協調動作を行うことにより、低遅延性と学習精度の向上を両立することを目指す。本研究で確立する技術は、自律協調制御の実装によって高度なAI処理を可能にする点で有用性が高く、他分野への応用も期待できる。

Outline of Final Research Achievements

Edge AI processing, which is characterized by low latency, is expected to be applied to self-driving vehicles and delivery drones, and is the subject of ongoing research and development. While edge AI processing can produce extremely low latency learning inference results, it has the problem of degrading learning inference accuracy because the processing is performed only on a single device with inferior performance. To solve this problem, we proposed an autonomous cooperative edge AI system in which multiple edge devices operate in a cooperative manner, and developed the basic technology for this system. With the proposed method, we have established one of the fundamental technologies that can bring out high performance as a whole by allowing devices with inferior individual performance to work together in an organic manner.

Academic Significance and Societal Importance of the Research Achievements

実行時に管理情報を交換しながらネットワーク的に近い位置のデバイスを動的にクラスタリングし、クラスタ間で自律的に処理を移送する制御が十分にシステム全体の潜在性能を引き出せることを示せた。これにより、エッジAI処理環境の新しい構成方式を示すことができた。本研究の成果は、複雑化するシステムを低オーバヘッドで自律協調制御する基盤技術になり得ることから、これまでこうした自律協調制御の適用が難しかったような他の分野に対しても応用が可能である。

Report

(4 results)
  • 2022 Annual Research Report   Final Research Report ( PDF )
  • 2021 Research-status Report
  • 2020 Research-status Report
  • Research Products

    (9 results)

All 2022 2021

All Journal Article (5 results) (of which Peer Reviewed: 5 results,  Open Access: 1 results) Presentation (4 results) (of which Int'l Joint Research: 4 results)

  • [Journal Article] Parallel Binary Search Tree Construction Inspired by Thread-Level Speculation2022

    • Author(s)
      Hirata Hiroaki、Nunome Atsushi
    • Journal Title

      Proceedings of the 23rd International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing

      Volume: 1 Pages: 74-81

    • DOI

      10.1109/snpd-summer57817.2022.00021

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Enhancing the Performance of an Autonomous Distributed Storage System in a Large-Scale Network2022

    • Author(s)
      Nunome Atsushi、Hirata Hiroaki
    • Journal Title

      Proceedings of the 23rd International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing

      Volume: 1 Pages: 87-94

    • DOI

      10.1109/snpd-summer57817.2022.00023

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Adaptive Parameter Tuning for Constructing Storage Tiers in an Autonomous Distributed Storage System2022

    • Author(s)
      Nunome Atsushi、Hirata Hiroaki
    • Journal Title

      International Journal of Networked and Distributed Computing

      Volume: 10 Issue: 1-2 Pages: 1-10

    • DOI

      10.1007/s44227-022-00004-3

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] An Adaptive Tiering Scheme for an Autonomous Distributed Storage System2021

    • Author(s)
      Atsushi Nunome, Hiroaki Hirata
    • Journal Title

      Proceedings of 8th ACIS International Virtual Conference on Applied Computing & Information Technology (ACIT 2021)

      Volume: -

    • Related Report
      2020 Research-status Report
    • Peer Reviewed
  • [Journal Article] Reducing the Repairing Penalty on Misspeculation in Thread-Level Speculation2021

    • Author(s)
      Hiroaki Hirata, Atsushi Nunome
    • Journal Title

      Proceedings of 8th ACIS International Virtual Conference on Applied Computing & Information Technology (ACIT 2021)

      Volume: -

    • Related Report
      2020 Research-status Report
    • Peer Reviewed
  • [Presentation] Parallel Binary Search Tree Construction Inspired by Thread-Level Speculation2022

    • Author(s)
      Hiroaki Hirata
    • Organizer
      23rd ACIS International Summer Virtual Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD 2022-Summer)
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Enhancing the Performance of an Autonomous Distributed Storage System in a Large-Scale Network2022

    • Author(s)
      Atsushi Nunome
    • Organizer
      23rd ACIS International Summer Virtual Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD 2022-Summer)
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] An Adaptive Tiering Scheme for an Autonomous Distributed Storage System2021

    • Author(s)
      Atsushi Nunome
    • Organizer
      8th ACIS International Virtual Conference on Applied Computing & Information Technology (ACIT 2021)
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research
  • [Presentation] Reducing the Repairing Penalty on Misspeculation in Thread-Level Speculation2021

    • Author(s)
      Hiroaki Hirata
    • Organizer
      8th ACIS International Virtual Conference on Applied Computing & Information Technology (ACIT 2021)
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research

URL: 

Published: 2020-04-28   Modified: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi