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

2021 Fiscal Year Annual Research Report

Comprehensive Analysis and Repairing Techniques for Stateful Deep Learning Systems

Research Project

Project/Area Number 20H04168
Research InstitutionKyushu University

Principal Investigator

馬 雷  九州大学, システム情報科学研究院, 准教授 (70842061)

Co-Investigator(Kenkyū-buntansha) 佐藤 周行  東京大学, 情報基盤センター, 准教授 (20225999)
趙 建軍  九州大学, システム情報科学研究院, 教授 (20299580)
石川 冬樹  国立情報学研究所, アーキテクチャ科学研究系, 准教授 (50455193)
鵜林 尚靖  九州大学, システム情報科学研究院, 教授 (80372762)
Project Period (FY) 2020-04-01 – 2023-03-31
Keywordsソフトウェア工学 / 知能ソフトウェア解析 / 深層学習 / 機械学習工学 / 回帰型ニューラルネットワーク解析 / 機械学習品質保証
Outline of Annual Research Achievements

深層学習(DL)は画像処理、音声認識、自然言語処理、及び囲碁などの応用面で華々しい成功をおさめ、自動運転車やロボットなど、社会基盤に関わる重要な分野での成果が益々期待されている。それと同時に、DLシステムにおける障害が発生する場合、社会に巨大な災害をもたらす可能性があるため、その信頼性と安全性に対する要求が益々高くなっている。本研究では、ステートフル深層学習システム(RNN)に対する系統的解析と修正技術の確立を目的としている。

今年度は昨年度の研究に基づいて、主要的な抽象モデルの抽出方法と解析とテスト生成フレームワークの構築と開発した。具体的には:1)RNNから抽象モデルの抽出方法とツールを完成し、提出された手法を実験で評価した。2) RNN自動解析とテスト生成フレームワークの開発した。3)RNNの修正技術を探索研究を行った。 4)それぞれの提出された技術が大規模で実験され、有効性を検証した。5)今年度それぞれの研究結果がまとめて国際会議と雑誌を投稿した。

Current Status of Research Progress
Current Status of Research Progress

1: Research has progressed more than it was originally planned.

Reason

当初提案されたの研究計画に向けて、順調に進んでいる。

Strategy for Future Research Activity

今後の研究実施計画では、2020~2021年度の研究に基づいて、主要的な目標はRNNテスト、分析と修正技術を提案と開発を完成し、同一フレームワークを開発と統合する。更に提案された技術とフレームワークは大規模の評価を行って、関連重要分野での実用性と適用性を検証する 。
具体的には: 1)カバレッジに基づくテストフレームワークの開発を統合と完成し、提出された手法を実験する。2)RNNのOnlineによる分析と修正技術を開発する。3)RNNのOfflineによる分析と修正技術を開発する。4)提出されたテスト、分析と修正技術が総合フレームワークで統合する。5) さらなる、各技術の実用性と適用性重要的な関連分野で大規模の検証と実験を行う、例えば:自然言語処理、音声認識など。6)それぞれ提出され技術と研究成果を執筆し、国際会議と雑誌論文を投稿する。

  • Research Products

    (20 results)

All 2022 2021 Other

All Int'l Joint Research (4 results) Journal Article (3 results) (of which Int'l Joint Research: 2 results,  Peer Reviewed: 3 results,  Open Access: 1 results) Presentation (13 results) (of which Int'l Joint Research: 13 results)

  • [Int'l Joint Research] Nanyang Technological University(シンガポール)

    • Country Name
      SINGAPORE
    • Counterpart Institution
      Nanyang Technological University
  • [Int'l Joint Research] University of Calgary(カナダ)

    • Country Name
      CANADA
    • Counterpart Institution
      University of Calgary
  • [Int'l Joint Research] Peking University/Tianjin University(中国)

    • Country Name
      CHINA
    • Counterpart Institution
      Peking University/Tianjin University
  • [Int'l Joint Research] Iowa State University(米国)

    • Country Name
      U.S.A.
    • Counterpart Institution
      Iowa State University
  • [Journal Article] NPC: Neuron Path Coverage via Characterizing Decision Logic of Deep Neural Networks2022

    • Author(s)
      Xie Xiaofei、Li Tianlin、Wang Jian、Ma Lei、Guo Qing、Juefei-Xu Felix、Liu Yang
    • Journal Title

      ACM Transactions on Software Engineering and Methodology

      Volume: 31 Pages: 1~27

    • DOI

      10.1145/3490489

    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Formalism-Driven Development: Concepts, Taxonomy, and Practice2022

    • Author(s)
      Ding Yepeng、Sato Hiroyuki
    • Journal Title

      Applied Sciences

      Volume: 12 Pages: 3415~3415

    • DOI

      10.3390/app12073415

    • Peer Reviewed
  • [Journal Article] DeepRepair: Style-Guided Repairing for Deep Neural Networks in the Real-World Operational Environment2021

    • Author(s)
      Yu Bing、Qi Hua、Qing Guo、Juefei-Xu Felix、Xie Xiaofei、Ma Lei、Zhao Jianjun
    • Journal Title

      IEEE Transactions on Reliability

      Volume: - Pages: 1~16

    • DOI

      10.1109/TR.2021.3096332

    • Peer Reviewed / Int'l Joint Research
  • [Presentation] NeuRecover: Regression-Controlled Repair of Deep Neural Networks with Training History2022

    • Author(s)
      Shogo Tokui, Susumu Tokumoto, Akihito Yoshii, Fuyuki Ishikawa, Takao Nakagawa, Kazuki Munakata and Shinji Kikuchi
    • Organizer
      The 29th IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER 2022), March 2022
    • Int'l Joint Research
  • [Presentation] Formalism-Driven Development of Decentralized Systems2022

    • Author(s)
      Yepeng Ding, Hiroyuki Sato
    • Organizer
      Proc. 26th Int'l Conf. Engineering of Complex Computer Systems, March, 2022
    • Int'l Joint Research
  • [Presentation] Learning to Adversarially Blur Visual Object Tracking2021

    • Author(s)
      Qing Guo, Ziyi Cheng, Felix Juefei-Xu, Lei Ma, Xiaofei Xie, Yang Liu, and Jianjun Zhao
    • Organizer
      International Conference on Computer Vision, Montreal, Canada, 2021 (ICCV 2021, CORE Rank A*)
    • Int'l Joint Research
  • [Presentation] Automatic RNN Repair via Model-based Analysis2021

    • Author(s)
      Xiaofei Xie, Wenbo Guo, Lei Ma, Wei Le, Jian Wang, Lingjun Zhou, Yang Liu, Xinyu Xing
    • Organizer
      The 38th International Conference on Machine Learning, 2021 (ICML 2021, CORE Rank A*)
    • Int'l Joint Research
  • [Presentation] Effective Hybrid System Falsification Using Monte Carlo Tree Search Guided by QB-Robustness2021

    • Author(s)
      Zhenya Zhang, Deyun Lyu, Paolo Arcaini, Lei Ma, Ichiro Hasuo and Jianjun Zhao
    • Organizer
      The 33rd International Conference on Computer-Aided Verification, 2021 (CAV 2021, CORE Rank A*)
    • Int'l Joint Research
  • [Presentation] Automatic Web Testing using Curiosity-Driven Reinforcement Learning2021

    • Author(s)
      Yan Zheng, Yi Liu, Xiaofei Xie, Yepang Liu, Lei Ma, Jianye Hao, and Yang Liu
    • Organizer
      The 43rd International Conference on Software Engineering (ICSE 2021, CORE Rank A*)
    • Int'l Joint Research
  • [Presentation] A Search-Based Testing Framework for Deep Neural Networks of Source Code Embedding2021

    • Author(s)
      Maryam V. Pour, Li Zhuo, Lei Ma and Hadi Hemmati
    • Organizer
      IEEE International Conference on Software Testing, Verification and Validation (ICST 2021, CORE Rank A)
    • Int'l Joint Research
  • [Presentation] Decision-Guided Weighted Automata Extraction from Recurrent Neural Networks2021

    • Author(s)
      Xiyue Zhang, Xiaoning Du, Xiaofei Xie, Lei Ma, Yang Liu, Meng Sun
    • Organizer
      Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI 2021, CORE Rank A*)
    • Int'l Joint Research
  • [Presentation] EfficientDeRain: Learning Pixel-wise Dilation Filtering for High-Efficiency Single-Image Deraining2021

    • Author(s)
      Qing Guo, Jingyang Sun, Felix Juefei-Xu, Lei Ma, Xiaofei Xie, Wei Feng, Yang Liu, and Jianjun Zhao
    • Organizer
      Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI 2021, CORE Rank A*)
    • Int'l Joint Research
  • [Presentation] What to Blame? On the Granularity of Fault Localization for Deep Neural Networks2021

    • Author(s)
      Matias Duran, Xiao-Yi Zhang, Paolo Arcaini, Fuyuki Ishikawa
    • Organizer
      The 32nd International Symposium on Software Reliability Engineering (ISSRE 2021 Practical Experience Reports), pp.264-275, October 2021
    • Int'l Joint Research
  • [Presentation] Model-based Data-Complexity Estimator for Deep Learning Systems2021

    • Author(s)
      Yuta Ojima, Shingo Horiuchi, Fuyuki Ishikawa
    • Organizer
      The 3rd IEEE International Conference on Artificial Intelligence Testing (IEEE AI Tests 2021), pp.1-8, August 2021
    • Int'l Joint Research
  • [Presentation] Dynamic resource allocation among collocated applications via reinforcement learning2021

    • Author(s)
      Li Shaowen, Li Gen, SATO Hiroyuki
    • Organizer
      Proc. IEEE 6th Int'l Conf. Cloud Computing and Big Data Analytics, A0014, Chengdu, April, 2021
    • Int'l Joint Research
  • [Presentation] In-air Signature Authentication Using Smartwatch Motion Sensors2021

    • Author(s)
      G. Li, L. Zhang and H. Sato
    • Organizer
      2021 IEEE 45th Annual Computers, Software, and Applications Conference (COMPSAC), 2021
    • Int'l Joint Research

URL: 

Published: 2022-12-28  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi