Seulbae Kim (김슬배)

[me] Ph.D.
School of Cybersecurity and Privacy (SCP), College of Computing
Georgia Institute of Technology
Advisor: Dr. Taesoo Kim

Research Interests

Cyber-physical systems security (drones, self-driving vehicles, robots, IoT); Test-driven security.

I am deeply passionate about investigating the security and robustness of Cyber-Physicsl Systems (CPS), which encompass practical systems such as drones, self-driving vehicles, robotic systems, and the Internet of Things (IoT) devices.

Throughout the journey of pursuing my M.S. and Ph.D. degrees, I have developed systematic testing methodologies tailored for each building block of CPS, including the OS layer (VUDDY [S&P'17], Hydra [SOSP'19]), communication protocol layer (under review), middleware layer (RoboFuzz [ESEC/FSE'22]), and application/logic layer (DriveFuzz [CCS'22]). These work have facilitated comprehensive testing of CPS, resulting in the discovery of critical bugs and vulnerabilities, as documented in the respective papers published in top-tier conferences in the fields of Security, Systems, and Software Engineering.

In my recently submitted work, I designed and implemented a fuzzer for a widely-used communication protocol in Cyber-Physical Systems. This fuzzer successfully uncovered numerous zero-day vulnerabilities, resulting in CVEs, and also identified specification violations within popular implementations of the protocol. Currently, the work is under review.

I am actively engaged in several fascinating projects at the moment. These include the preemptive physical security measures for robotic vehicles, automated causality analysis of robotic bugs, and the development of robust attack detection mechanisms under varying physical properties, to name just a few.

Publications

  1. DriveFuzz: Discovering Autonomous Driving Bugs through Driving Quality-Guided Fuzzing
    Seulbae Kim, Major Liu, Junghwan Rhee, Yuseok Jeon, Yonghwi Kwon, and Chung Hwan Kim
    CCS'22: In Proceedings of the 29th ACM Conference on Computer and Communications Security
    Los Angeles, USA, Nov 2022     [ paper ] [ extended paper ] [ slides ] [ code ] [ website ] [ news ]
  2. RoboFuzz: Fuzzing Robotic Systems over Robot Operating System (ROS) for Finding Correctness Bugs
    Seulbae Kim, and Taesoo Kim
    ESEC/FSE'22: In Proceedings of the 30th ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering
    Singapore, Nov 2022     [ paper ] [ slides ] [ code ]
  3. Riding the IoT Wave with VFuzz: Discovering Security Flaws in Smart Home
    Carlos Nkuba, Seulbae Kim, Sven Dietrich, and Heejo Lee
    IEEE Access, Vol. 10
    December 2021     [ paper ] [ code ] [ CVEs ]
  4. CENTRIS: A Precise and Scalable Approach for Identifying Modified Open-Source Software Reuse
    Seunghoon Woo, Sunghan Park, Seulbae Kim, Heejo Lee, and Hakjoo Oh
    ICSE'21: In Proceedings of the 43rd International Conference on Software Engineering
    Virtual, May 2021     [ paper ] [ slides ] [ code ] [ news 1, news 2 (Kor) ]
  5. Finding Bugs in File Systems with an Extensible Fuzzing Framework
    Seulbae Kim, Meng Xu, Sanidhya Kashyap, Jungyeon Yoon, Wen Xu, and Taesoo Kim
    ACM Transcations on Storage, Vol. 16, No.2, Article 10
    May 2020     [ paper ] [ ACM Library ]
  6. Finding Semantic Bugs in File Systems with an Extensible Fuzzing Framework
    Seulbae Kim, Meng Xu, Sanidhya Kashyap, Jungyeon Yoon, Wen Xu, and Taesoo Kim
    SOSP'19: In Proceedings of the 27th ACM Symposium on Operating Systems Principles
    Ontario, Canada, October 2019     [ paper ] [ slides ] [ presentation ] [ code ] [ bugs ] [ news 1, news 2 ]
  7. Software systems at risk: An empirical study of cloned vulnerabilities in practice
    Seulbae Kim, and Heejo Lee
    Computers & Security, Vol. 77
    August 2018     [ paper ]
  8. VUDDY: A Scalable Approach for Vulnerable Code Clone Discovery
    Seulbae Kim, Seunghoon Woo, Heejo Lee, and Hakjoo Oh
    S&P'17: In Proceedings of the 38th IEEE Symposium on Security and Privacy
    San Jose, CA, May 2017     [ paper ] [ slides ] [ presentation ] [ code ] [ news 1 (Kor), news 2 (Kor) ]
  9. SIGMATA: Storage Integrity Guaranteeing Mechanism against Tampering Attempts for Video Event Data Recorders
    Hyuckmin Kwon, Seulbae Kim, and Heejo Lee
    IMCIC 2016: In Proceedings of the 7th International Multi-conference on Complexity, Informatics, and Cybernatics
    Orlando, FL, March 2016.
    Session's best paper award     [ paper ] [ slides ]

Education

  1. Georgia Institute of Technology Aug 2018 - Dec 2023
    Ph.D. in Computer Science, College of Computing
  2. Korea University Mar 2016 - Aug 2018
    M.S. in Computer Science and Engineering
    Thesis: Scalable Approach for Code Clone Detection and its Application in Practice
    Advisor: Dr. Heejo Lee
  3. Korea University Mar 2010 - Feb 2016
    B.S. in Computer Science and Engineering
  4. Goyang Foreign Language High School Mar 2007 - Feb 2010
    Major: English, Track: Natural Science

Teaching

Services

Last updated: Nov 21, 2023