I am currently a Research Fellow at National Institute of Informatics, working in the Trustworthy & Smart Software Engineering Lab led by Prof. Fuyuki ISHIKAWA. Before that, I received my PhD from Department of Information Science and Electrical Engineering (ISEE), Kyushu University. My supervisor was Prof. Jianjun Zhao, and I was co-advised by Prof. Paolo Arcaini and Prof. Zhenya Zhang. My research interests mainly focus on quality assurance of AI-enabled cyber-physical systems (AI-enabled CPSs). Research Interests (including but not limited to): SE4AI, Formal Methods, Trustworthy Cyber-Physical Systems, Safe Autonomous Driving Systems I am eager to exchange ideas and explore collaborations with researchers who share these interests. I am also seeking new research and career opportunities. Email: lyudeyun@nii.ac.jp [Google Scholar][Github] [LinkedIn][Curriculem Vitae]

at Tokyo, Japan. (Dec 29, 2020)

News

• [2025-08-05] Honored with the 2025 Distinguished Research Award from IPSJ/SIGSE (2025年度情報処理学会ソフトウェア工学研究会卓越研究賞).

• [2025-06-27] My PhD dissertation “Quality Assurance of AI-Enabled Cyber-Physical Systems” was selected for the 2025 IPSJ Doctoral Dissertation Recommendation.

• [2025-04-23] “Fault Localization of AI-Enabled Cyber-Physical Systems by Exploiting Temporal Neuron Activation” is accepted at JSS 2025.

• [2025-04-15] Join NII as a Research Fellow.

• [2025-03-05] “Automated Generation of Benchmarks for Falsification of STL Specifications” is accepted at TCAD 2025.

• [2024-11-03] “SpectAcle: Fault Localisation of AI-Enabled CPS by Exploiting Sequences of DNN Controller Inferences” is accepted at TOSEM 2024.

• [2024-09-11] Give an oral presentation at the Top Conference/Journal Special Talk (トップカンファレンス・トップ論文誌特別講演), JSSST 2024.

• [2024-06-13] Join AAAI 2025 as a PC Member.

• [2024-03-22] “Search-Based Repair of DNN Controllers of AI-Enabled Cyber-Physical Systems Guided by System-Level Specifications” is accepted at GECCO 2024.

• [2023-09-18] Join the National Institute of Informatics for a six-month research internship, supervised by Prof. Fuyuki ISHIKAWA and Paolo Arcaini.

• [2023-07-13] Join AAAI 2024 as a PC Member.

• [2023-02-08] “FalsifAI: Falsification of AI-Enabled Hybrid Control Systems Guided by Time-Aware Coverage Criteria” is accepted at ICSE 2023 as Journal-First Paper.

• [2022-12-22] “Towards Building Reliable AI-Enabled Cyber-Physical Systems” is accepted at the Doctoral Symposium, FM 2023.

• [2022-08-31] Give a poster presentation at the Demo/Poster Session (デモ・ポスターセッション), JSSST 2022.

• [2022-08-10] Join AAAI 2023 as a PC Member.

• [2022-07-21] “FalsifAI: Falsification of AI-Enabled Hybrid Control Systems Guided by Time-Aware Coverage Criteria” is accepted at TSE 2022.

• [2022-01-11] “When Cyber-Physical Systems Meet AI: A Benchmark, an Evaluation, and a Way Forward” is accepted at ICSE SEIP 2022.

• [2022-01-08] “An Verification Framework for Behavioral Safety of Self-Driving Cars” is accepted at IET Intelligent Transport System 2022.

• [2021-08-20] “Time-Aware Coverage Criteria for Testing of AI-Enabled Hybrid Control Systems” is accepted at ASE LBR 2021.

• [2021-04-20] “SDLV: Verification of Steering Angle Safety for Self-Driving Cars” is accepted at FAC 2021.

• [2021-04-18] “Effective Hybrid System Falsification Using Monte Carlo Tree Search Guided by QB-Robustness” is accepted at CAV 2021.

• [2021-02-24] “On the Effectiveness of Signal Rescaling in Hybrid System Falsification” is accepted at NFM 2021.

Experience

• Research Fellow at National Institute of Informatics, Apr. 2025 - Present

• Ph.D. in Advanced Information Technology, Kyushu University, Oct. 2020 - Mar. 2025

• M.S. in Software Engineering, Dalian University of Technology, Sept. 2018 - Jun. 2020

• B.S. in Network Engineering, Dalian University of Technology, Sept. 2014 - Jun. 2018

Publications († denotes co-first authorship)

Conference

• Search-Based Repair of DNN Controllers of AI-Enabled Cyber-Physical Systems Guided by System-Level Specifications [PDF] [Code]
  Deyun Lyu, Zhenya Zhang, Paolo Arcaini, Fuyuki Ishikawa, Thomas Laurent, Jianjun Zhao
  The Genetic and Evolutionary Computation Conference (GECCO 2024)

• When Cyber-Physical Systems Meet AI: A Benchmark, an Evaluation, and a Way Forward
  (SEIP - Software Engineering in Practice)
  Jiayang Song†, Deyun Lyu†, Zhenya Zhang, Zhijie Wang, Tianyi Zhang, Lei Ma
  The 44th International Conference on Software Engineering (ICSE SEIP 2022)

• Effective Hybrid System Falsification Using Monte Carlo Tree Search Guided by QB-Robustness
  Zhenya Zhang, Deyun Lyu, Paolo Arcaini, Lei Ma, Ichiro Hasuo, Jianjun Zhao
  The 33th International Conference on Computer-Aided Verification (CAV 2021)

• On the Effectiveness of Signal Rescaling in Hybrid System Falsification [Code]
  Zhenya Zhang, Deyun Lyu, Paolo Arcaini, Lei Ma, Ichiro Hasuo, Jianjun Zhao
  The 13th NASA Formal Methods Symposium (NFM 2021)

Journal

• Fault Localization of AI-Enabled Cyber-Physical Systems by Exploiting Temporal Neuron Activation [PDF]
  Deyun Lyu, Yi Li, Zhenya Zhang, Paolo Arcaini, Xiao-Yi Zhang, Fuyuki Ishikawa, Jianjun Zhao
  Journal of Systems and Software, 2025.

• Automated Generation of Benchmarks for Falsification of STL Specifications
  Yipei Yan, Deyun Lyu, Zhenya Zhang, Paolo Arcaini, Jianjun Zhao
  IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2025.

• SpectAcle: Fault Localisation of AI-Enabled CPS by Exploiting Sequences of DNN Controller Inferences [PDF]
  Deyun Lyu, Zhenya Zhang, Paolo Arcaini, Xiao-Yi Zhang, Fuyuki Ishikawa, Jianjun Zhao
  ACM Transactions on Software Engineering and Methodology, 2024.

• FalsifAI: Falsification of AI-Enabled Hybrid Control Systems Guided by Time-Aware Coverage Criteria
  Zhenya Zhang, Deyun Lyu, Paolo Arcaini, Lei Ma, Ichiro Hasuo, Jianjun Zhao
  IEEE Transactions on Software Engineering, 2023, 49(4): 1842 - 1859.

• An Verification Framework for Behavioral Safety of Self-Driving Cars
  Huihui Wu, Deyun Lyu, Yanan Zhang, Gang Hou, Masahiko Watanabe, Jie Wang, Weiqiang Kong
  IET Intelligent Transport Systems, 2022, 16(5): 630-647.

• SDLV: Verification of Steering Angle Safety for Self-Driving Cars
  Huihui Wu, Deyun Lyu, Tengxiang Cui, Gang Hou, Masahiko Watanabe, Weiqiang Kong
  Formal Aspects of Computing, 2021, 33, 325-341.

Workshop & Presentation & Poster

• Search-Based Repair of DNN Controllers of AI-Enabled Cyber-Physical Systems Guided by System-Level Specifications [PDF]
  Top Conference/Journal Special Talk (トップカンファレンス・トップ論文誌特別講演), 日本ソフトウェア科学会第41回大会 (JSSST 41), 2024, Osaka, Japan.

• Towards Building Reliable AI-Enabled Cyber-Physical Systems [PDF]
  Doctoral Symposium, The 26th International Symposium on Formal Methods (FM 2023).

• FalsifAI: Falsification of AI-Enabled Hybrid Control Systems Guided by Time-Aware Coverage Criteria [PDF]
  Demo/Poster Session (デモ・ポスターセッション), 日本ソフトウェア科学会第39回大会 (JSSST 39), 2022, Nagoya, Japan.

Patent

• Method for adaptively detecting chessboard sub-pixel level corner points
  U.S. Patent No.: 11,900,634.
  Name(s) of Inventor(s): Wei Zhong, Deyun Lyu, Weiqiang Kong, Risheng Liu, Xin Fan, Zhongxuan Luo, Shengquan Li.
  

• Method for fully automatically detecting chessboard corner points
  U.S. Patent No.: 12,094,152.
  Name(s) of Inventor(s): Weiqiang Kong, Deyun Lyu, Wei Zhong, Risheng Liu, Xin Fan, Zhongxuan Luo.
  

• Method for infrared small target detection based on depth map in complex scene
  U.S. Patent No.: 12,108,022.
  Name(s) of Inventor(s): Weiqiang Kong, Deyun Lyu, Wei Zhong, Risheng Liu, Xin Fan, Zhongxuan Luo.
  

Research Experiences

• Visiting Researcher at National Institute of Informatics, Sep 2023 - Feb 2024, Tokyo, Japan
  1. DNN Repair of AI-enabled cyber-physical systems
  Work with Paolo Arcaini and Fuyuki Ishikawa, supported by the MIRAI-eAI Project.

• SPRING Research Fellow, Kyushu University, Oct 2021 - Sep 2023, Fukuoka, Japan
  

• Junior Research Engineer at North Automatic Control Technology Institute, Jun 2018 - Sep 2019, Taiyuan, China
  1. Corner Detection
  Present an efficient checkerboard corner full-automatic detection method, using traditional techniques such as binarization, edge detection.
  2. Camera Calibration
  Present an efficient method for optimizing external parameters of binocular camera.
  3. Target Detection
  Persent an infrared small target detection method based on depth map in complex scene.
  

Professional Services

• Chair:
  Web Chair, FM 2026
  

• PC Member:
  ATVA 2025 Artifact Evaluation, AAAI 2025, AAAI 2024, AAAI 2023
  

• Reviewer:
  TOSEM 2025, IROS 2025, ICECCS 2024, ICSE 2021, ICML 2022
  

• Shadow Reviewer:
  TSE, TOSEM, ICSE, ASE, FSE, ISSTA, EMSOFT, ICST, TASE, EMSE, TR, CAIN ……

Teaching Assistant Experiences

• Python Programming Exercises (Pythonプログラミング演習) (for undergraduate students), Summer 2024

• Python Programming Exercises (Pythonプログラミング演習) (for undergraduate students), Summer 2023

• Machine Learning Systems Engineering (for graduate students), Summer 2021

Awards

• 2025 Distinguished Research Award from IPSJ/SIGSE (2025年度情報処理学会ソフトウェア工学研究会卓越研究賞)

• Ph.D. dissertation selected as a 2025 IPSJ SIG-SE Recommended Doctoral Thesis (研究会推薦博士論文), Information Processing Society of Japan

• 次世代研究者挑戦的研究プログラム, Supported by the Japan Science and Technology Agency (JST)
  

• Kyushu University Graduate Research Support Scholarship (九州大学大学院研究支援奨学金)

Conference Travel Awards

• GECCO Student Travel Grant ($600 USD), The Genetic and Evolutionary Computation Conference (GECCO 2024)
  

Visitors