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

Fundamental Technologies of Efficient Analysis and Design for Approximate Computing-based Embedded Systems

Research Project

Project/Area Number 23K20377
Project/Area Number (Other) 20H04154 (2020-2023)
Research Category

Grant-in-Aid for Scientific Research (B)

Allocation TypeMulti-year Fund (2024)
Single-year Grants (2020-2023)
Section一般
Review Section Basic Section 60040:Computer system-related
Research InstitutionInstitute of Science Tokyo

Principal Investigator

原 祐子  東京工業大学, 工学院, 准教授 (20640999)

Project Period (FY) 2020-04-01 – 2025-03-31
Project Status Granted (Fiscal Year 2024)
Budget Amount *help
¥17,420,000 (Direct Cost: ¥13,400,000、Indirect Cost: ¥4,020,000)
Fiscal Year 2024: ¥2,730,000 (Direct Cost: ¥2,100,000、Indirect Cost: ¥630,000)
Fiscal Year 2023: ¥3,640,000 (Direct Cost: ¥2,800,000、Indirect Cost: ¥840,000)
Fiscal Year 2022: ¥3,640,000 (Direct Cost: ¥2,800,000、Indirect Cost: ¥840,000)
Fiscal Year 2021: ¥3,900,000 (Direct Cost: ¥3,000,000、Indirect Cost: ¥900,000)
Fiscal Year 2020: ¥3,510,000 (Direct Cost: ¥2,700,000、Indirect Cost: ¥810,000)
Keywords近似計算 / 乱数生成器 / 組込みシステム / Internet of Things (IoT) / 組み込みシステム / Approximate Computing
Outline of Research at the Start

Internet of Things (IoT) デバイスの発展に伴い、益々膨大なデータが生成される中、IoT/組込みシステム上で即時にデータ処理するエッジコンピューティング技術の確立は、学術研究発展・市場開拓にとって喫緊の課題である。IoTアプリケーションの多くは、多少の計算誤差を許容可能な特徴があり、近似計算という新しいデータ処理技術は、上記の課題の大きなブレークスルーになる。本研究は、IoTアプリケーションを適切に近似計算するための基礎技術を確立し、近似計算の学術基盤と効率的な利活用技術を確立する。

Outline of Annual Research Achievements

Internet of Things (IoT) デバイスの発展に伴い、益々膨大なデータが生成される中、IoT/組込みシステム上で即時にデータ処理するエッジコンピューティング技術の確立は、学術研究発展・市場開拓にとって喫緊の課題である。IoTアプリケーションの多くは、多少の計算誤差を許容可能な特徴があり、近似計算という新しいデータ処理技術は、上記の課題の大きなブレークスルーになる。本研究は、IoTアプリケーションを適切に近似計算するための基礎技術を確立する。特に「どの程度」と「どのぐらいの頻度で」という2つの大きな問いに答えるため、近似計算の学術基盤と効率的な利活用技術を確立する。研究代表者は、前年度、これまでの研究成果によって得られた知見を有効活用するAC手法を整理し、不足する部分については新たなAC手法の開発を行った。特に、効率的なプログラム解析を実現するため、Graph Neural Networkを取り入れる方法に着手した。
本年度は、前年度に引き続き、Graph Neural Networkを用いたプログラム解析手法の実装を進めた。具体的には、ある制約の下で、最も類似性の高いプログラムを部分的に抽出する手法に取り組んだ。さらに、実用的なアプリケーション(様々な方式の機械学習および暗号システムの乱数生成器等)に対してデータ及びアルゴリズム由来の近似化耐性および近似化の影響(ハードウェアコスト)の調査・評価を行った。

Current Status of Research Progress
Current Status of Research Progress

2: Research has progressed on the whole more than it was originally planned.

Reason

Graph Neural Networkを取り入れた画期的なプログラムの解析手法の実装を進めた。また、暗号システムのプリミティブなコンポーネントである乱数生成器に対して、近似計算を応用する画期的な実装方法に取り組み、問題点や知見を明らかにした。

Strategy for Future Research Activity

更に最先端の乱数生成器を対象に、近似化耐性の調査・評価を行う。さらに、最終年度は本研究期間中に試行した様々なアプリケーションへのソフトウェアおよびハードウェアのAC手法に関する問題と知見を整理する。

Report

(4 results)
  • 2023 Annual Research Report
  • 2022 Annual Research Report
  • 2021 Annual Research Report
  • 2020 Annual Research Report
  • Research Products

    (25 results)

All 2024 2023 2022 2021 2020

All Journal Article (3 results) (of which Int'l Joint Research: 1 results,  Peer Reviewed: 3 results,  Open Access: 3 results) Presentation (22 results) (of which Int'l Joint Research: 14 results)

  • [Journal Article] Hardware/Software Codesign of Real-time Intrusion Detection System for Internet of Things Devices2024

    • Author(s)
      Qingyu Zeng and Yuko Hara-Azumi
    • Journal Title

      IEEE Internet of Things Journal

      Volume: - Issue: 12 Pages: 22351-22363

    • DOI

      10.1109/jiot.2024.3380822

    • Related Report
      2023 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Real-time Reconfiguration of Time-Aware Shaper for ULL Transmission in Dynamic Conditions2021

    • Author(s)
      Yu Nakayama, Ryo Yaegashi, Anh Hoang Ngoc Nguyen, Yuko Hara-Azumi
    • Journal Title

      IEEE Access

      Volume: 9 Pages: 115246-115255

    • DOI

      10.1109/access.2021.3105420

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] APNAS: Accuracy-and-Performance-Aware Neural Architecture Search Considering Neural Hardware Accelerators2020

    • Author(s)
      Paniti Achararit, Muhammad Abdullah Hanif, Rachmad Vidya Wicaksana Putra, Muhammad Shafique, Yuko Hara-Azumi
    • Journal Title

      IEEE Access

      Volume: 8 Pages: 165319-165334

    • DOI

      10.1109/access.2020.3022327

    • Related Report
      2020 Annual Research Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Presentation] Tram-FL: Routing-based Model Training for Decentralized Federated Learning2024

    • Author(s)
      Kota Maejima, Takayuki Nishio, Asato Yamazaki, Yuko Hara-Azumi
    • Organizer
      Consumer Communications & Networking Conference (CCNC)
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] eFPGAを用いた組込みプロセッサのIP保護2024

    • Author(s)
      一岡 知佑, タンビア アーメド, 原 祐子
    • Organizer
      ハードウェアセキュリティ研究会
    • Related Report
      2023 Annual Research Report
  • [Presentation] 故障注入攻撃耐性を強化した擬似乱数生成器の設計2024

    • Author(s)
      角 颯太, 楊 明宇, 原 祐子
    • Organizer
      ハードウェアセキュリティ研究会
    • Related Report
      2023 Annual Research Report
  • [Presentation] Graph Neural Networkを用いた組み込みプロセッサ拡張2024

    • Author(s)
      卯木 あゆ美, 原 祐子
    • Organizer
      VLSI設計技術研究会
    • Related Report
      2023 Annual Research Report
  • [Presentation] TI-AESに使用する擬似乱数生成器の物理安全性への影響2024

    • Author(s)
      原田 優咲, 塚原 麻輝, 宮原 大輝, 李 陽, 原 祐子, 崎山 一男
    • Organizer
      暗号と情報セキュリティシンポジウム (SCIS)
    • Related Report
      2023 Annual Research Report
  • [Presentation] A Highly Efficient Reinforcement Learning Based DFG Mapping Method on CGRA2023

    • Author(s)
      Tanvir Ahmed and Yuko Hara-Azumi
    • Organizer
      the Work-in-Progress (WiP) session of Design Automation Conference (DAC)
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Convergence Improvement by Parameters Exchange in Asynchronous Decentralized Federated Learning for Non-IID Data2023

    • Author(s)
      Asato Yamazaki, Takayuki Nishio, and Yuko Hara-Azumi
    • Organizer
      Euromicro Conference on Software Engineering and Advanced Applications (SEAA)
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Dynamic Split Computing-Aware Mixed-Precision Quantization for Efficient Deep Edge Intelligence2023

    • Author(s)
      Naoki Nagamatsu and Yuko Hara-Azumi
    • Organizer
      International Conference on Embedded and Ubiquitous Computing
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] 敵対的攻撃に対して頑健な軽量CNNモデルの自動探索2023

    • Author(s)
      小野 悠真, パニティ アッシャラリット, 原 祐子
    • Organizer
      暗号と情報セキュリティシンポジウム (SCIS)
    • Related Report
      2022 Annual Research Report
  • [Presentation] Hardware SAT Solver-based Area-efficient Accelerator for Autonomous Driving2022

    • Author(s)
      Yusuke Inuma and Yuko Hara-Azumi
    • Organizer
      International Conference on Field-Programmable Technology (ICFPT)
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Skip & Swap: Efficient Weight Spreading for Decentralized Machine Learning with Non-IID Data2022

    • Author(s)
      Asato Yamazaki, Takayuki Nishio, and Yuko Hara-Azumi
    • Organizer
      Work-in-Progress of Asia Pacific Conference on Robot IoT System Development and Platform (APRIS)
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] A Self-Attention Network for Deep JSCCM: The Design and FPGA Implementation2022

    • Author(s)
      Shohei Fujimaki, Yoshiaki Inoue, Daisuke Hisano, Kazuki Maruta, Yu Nakayama, and Yuko Hara-Azumi
    • Organizer
      IEEE Global Communications Conference (GlobeCOM)
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Gossip Swap SGD: Lightweight Decentralized Machine Learning for Non-homogeneous Data Distribution2022

    • Author(s)
      Naoya Yokota and Yuko Hara-Azumi
    • Organizer
      International Conference on Intelligent Software Methodologies, Tools, and Techniques
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Examining Vulnerability of HLS-designed Chaskey-12 Circuits to Power Side-Channel Attacks2022

    • Author(s)
      Saya Inagaki, Mingyu Yang, Yang Li, Kazuo Sakiyama and Yuko Hara-Azumi
    • Organizer
      International Symposium on Quality Electronic Design (ISQED)
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] 高位合成を用いたGentleman-Sande型NTTの高速化2022

    • Author(s)
      佐々木 智大, 原 祐子
    • Organizer
      第199回システムとLSIの設計技術研究発表会
    • Related Report
      2022 Annual Research Report
  • [Presentation] Neural Architecture SearchによるSplit Inferenceの効率化2022

    • Author(s)
      永松 直樹, 原 祐子
    • Organizer
      第199回システムとLSIの設計技術研究発表会
    • Related Report
      2022 Annual Research Report
  • [Presentation] Edge Domain Adaptation through Stepwise Cross-Domain Distillation2021

    • Author(s)
      Taisei Yamana, Yuko Hara-Azumi
    • Organizer
      Asia Pacific Conference on Robot IoT System Development and Platform (APRIS)
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Weight Exchange in Decentralized Distributed Machine Learning for Resource-Constrained IoT Edges2021

    • Author(s)
      Naoya Yokota, Yuko Hara-Azumi
    • Organizer
      Asia Pacific Conference on Robot IoT System Development and Platform (APRIS)
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] An FPGA-based Stochastic SAT Solver Leveraging Inter-Variable Dependencies2021

    • Author(s)
      Anh Hoang Ngoc Nguyen, Yuko Hara-Azumi
    • Organizer
      International Conference on Field-Programmable Logic and Applications (FPL)
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Zero Correlation Error: A Metric for Finite-Length BitstreamIndependence in Stochastic Computing2021

    • Author(s)
      Hsuan Hsiao, Joshua San Miguel, Yuko Hara-Azumi, Jason H. Anderson
    • Organizer
      Asia and South Pacific Design Automation Conference (ASP-DAC)
    • Related Report
      2020 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Amoeba-inspired System Controller on IoT Edge2021

    • Author(s)
      Anh Nguyen, Yuko Hara-Azumi
    • Organizer
      Ph.D. Forum of Design, Automation & Test in Europe (DATE)
    • Related Report
      2020 Annual Research Report
    • Int'l Joint Research
  • [Presentation] FPGA-SATソルバを用いたGate Control Listの高速アップデート法2021

    • Author(s)
      八重樫 遼, 原 祐子, Anh Hoang Ngoc Nguyen, 中山 悠
    • Organizer
      電子情報通信学会総合大会
    • Related Report
      2020 Annual Research Report

URL: 

Published: 2020-04-28   Modified: 2024-12-25  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi