photo

Aaron Wu   吴亚伦

I am a third-year Ph.D. student at the Beijing Key Laboratory of Security and Privacy in Intelligent Transportation at Beijing Jiaotong University (BJTU). I am affiliated with the Laboratory of Thorough Evaluation of Threats in the Artificial Intelligence (THETA Lab) and supervised by Prof. Zhen Han, Prof. Wenjia Niu and Dr. Endong Tong.

My research interests in AI security and CPS/IoT Security. I have authored or co-authored about 20 academic papers in international conferences and journals, including TAI, TGCN, TST, TICPS, IJIS, IEEE Network, etc. I’m open to any kind of collaboration. If you are interested to chat with me, please feel free to contact me through email.

  wuyalun1 [at] bjtu.edu.cn             Google Scholar             Github

Research Interests

Recently, I have been interested in exploring various aspects of AI security, including the reliability and robustness of machine learning models, data security and privacy protection in big data environments, and the decision control of reinforcement learning and intelligent agent robots. Our goal is to reveal the vulnerabilities of these models and raise awareness, advocating for increased research dedicated to AI security. The specific intersts are listed below:

1) CPS/IoT security:

Attacks and defenses against the physical world, sensors, algorithms and intelligent systems.

2) Data Augmentation:

Image/Instance generation, especially focus on generative data augmentation (GDA) for addressing corner cases and long-tail problems.

3) Backdoor Learning:

Backdoor attacks and defenses on deep neural networks (DNNs).

4) Reinforcement Learning:

Safety and robustness in reinforcement learning for robots and self-driving systems.

News

  • [2024/10] Accepted! One paper has been accepted by CCF CIVS 2024 and won the Best Student Paper Award. Congratulations!
  • [2024/09] Accepted! One paper has been accepted by International Journal of Intelligent Systems (IJIS).
  • [2024/09] Accepted! One paper has been accepted by ADMA 2024.
  • [2024/02] Accepted! One paper has been accepted by IEEE Network.
  • [2022/08] Our team won the Global Runner-up in the DEF CON'30 Autonomous Driving CTF Challenge. Checkout the news from some of the winning teams: [Our Team@BJTU], [Whitezard@FUDAN]. Congratulations!
  • [2022/06] I obtained a Master’s degree in Software Engineering from Beijing Jiaotong University (BJTU).
  • [2022/03] Our team won the First Price in the 2022 Beijing Winter Olympics Vulnerability Discovery Competition. Congratulations! [News]
  • [2022/02] I received a letter of appreciation from the National Major Network Security Test and Assessment Activity. [News]
  • [2021/08] Our team won the Global Runner-up in the DEFCON'29 Autonomous Driving CTF Challenge. Checkout the news from some of the winning teams: [Our Team@BJTU], [Cacti@UB], [Redbud@THU]. Congratulations!
  • [2021/05] Accepted! One paper has been accepted by ICICS 2021.

Selected Publications

All publications can be found on my Google Scholar page.

RetroreflectionBA: Leveraging Retroreflection as a Backdoor Attack Trigger for Fooling Pedestrian Detection Models.
Qiong Li, Yalun Wu, Qihuan Li, Xiaoshu Cui, Xiaolin Chang, Peng Sun, Jiqiang Liu, Wenjia Niu.
China Computer Federation Intelligent Vehicles Symposium (CCF CIVS 2024)
Best Student Paper Award [Paper] [PDF] [BibTex]

Collaborative Attack Sequence Generation Model Based on Multi-Agent Reinforcement Learning for Intelligent Traffic Signal System.
Yalun Wu*, Yingxiao Xiang*, Thar Baker, Endong Tong, Ye Zhu, Xiaoshu Cui, Zhenguo Zhang, Zhen Han, Jiqiang Liu, Wenjia Niu.
International Journal of Intelligent Systems (IJIS 2024)
(*Equal Contribution) [Paper] [PDF] [BibTex]

Nightfall Deception: A Novel Backdoor Attack on Traffic Sign Recognition Models via Low-Light Data Manipulation.
Yalun Wu, Qiong Li, YingXiao Xiang, Jinkai Zheng, Xingyu Wu, Zhen Han, Jiqiang Liu, Wenjia Niu.
International Conference on Advanced Data Mining and Applications (ADMA 2024)
[Paper] [PDF] [BibTex]

Lurking in the Shadows: Imperceptible Shadow Black-Box Attacks against Lane Detection Models.
Xiaoshu Cui, Yalun Wu, Yanfeng Gu, Qiong Li, Endong Tong, Jiqiang Liu, Wenjia Niu.
International Conference on Knowledge Science, Engineering and Management (KSEM 2024)
[Paper] [PDF] [BibTex]

Improving the Robustness of Pedestrian Detection in Autonomous Driving with Generative Data Augmentation.
Yalun Wu, Yingxiao Xiang, Endong Tong, Yuqi Ye, Zhibo Cui, Yunzhe Tian, Lejun Zhang, Jiqiang Liu, Zhen Han, Wenjia Niu.
IEEE Network 2024
[Paper] [PDF] [BibTex]

Blockchain-based decentralized federated learning: A secure and privacy-preserving system.
Shuang Zhao, Yalun Wu, Rui Sun, Xiaoai Qian, Dong Zi, Zhiqiang Xie, Endong Tong, Wenjia Niu, Jiqiang Liu, Zhen Han.
IEEE International Conference on High Performance Computing and Communications (HPCC 2021)
[Paper] [PDF] [BibTex]

Improving Convolutional Neural Network-Based Webshell Detection Through Reinforcement Learning.
Yalun Wu, Minglu Song, Yike Li, Yunzhe Tian, Endong Tong, Wenjia Niu, Bowei Jia, Haixiang Huang, Qiong Li, Jiqiang Liu.
International Conference on Information and Communications Security (ICICS 2021)
Acceptance Rate: 49/202≈24.25% [Paper] [PDF] [BibTex]

Selected Awards

  • Advanced Class Award, Beijing Municipal Commission of Education, 2023. (Team Leader)
  • Advanced Class of Graduate Students Award, Beijing Jiaotong University, 2023. (Team Leader)
  • XiaoMi Scholarship, Xiaomi Foundation, 2023. Thanks for Xiaomi Foundation and Xiaomi Corporation!
  • Excellent Scholarship in Social Work, Beijing Jiaotong University, 2023.
  • Excellent Graduate Student Cadre, Beijing Jiaotong University, 2023.
  • Excellent Part-time Counselor, Beijing Jiaotong University, 2022.
  • Global Runner-up in the DEF CON'30 Autonomous Driving CTF Challenge, Baidu Security & UCI, 2022. (Team Leader) [News]
  • First Price of the 2022 Beijing Winter Olympics Vulnerability Mining Competition, Peng Cheng Laboratory, 2022. (Team Leader) [News]
  • A letter of appreciation from the National Major Network Security Test and Assessment Activity, Peng Cheng Laboratory, 2022. (Team Leader) [News]
  • Global Runner-up in the DEF CON'29 Autonomous Driving CTF Challenge, Baidu Security & UCI, 2021. (Team Leader) [News]