Shuofeng Liu

Shuofeng Liu

Second-year Ph.D. student at School of EECS, The University of Queensland

About Me

I am currently a Ph.D. student at the School of Electrical Engineering and Computer Science, the University of Queensland, supervised by A/Prof. Guangdong Bai and Dr. Jason Xue. My research focuses on tackling real-world security & privacy issues, especially on building trustworthy AI and designing privacy-preserving systems.

Research Interests

  • Trustworthy AI
    • Modifier-based jailbreaking on T2I models: ModX@Oakland'25
    • IP Vulnerabilites on Distributed ML: MoEx@CIKM'24
    • Robustness Formalization for Language Model@ICFEM'23
  • Privacy compliance
    • Algorithmic Purpose Limitation: AlgoSpec@USENIX Security'24
    • LSTM-based Water Level Prediction Method: Hybrid@JASSE

News

  • [Mar. 2025] Our paper on modifier-based jailbreaking on text-to-image models is accepted by Oakland’25!
  • [Nov. 2024] I become an artificat evaluation (AE) reviewer of USENIX Security’25!
  • [Sep. 2024] I pass my Ph.D. milestone 1 progress review (confirmation)!
  • [Sep. 2024] I am invited to become a reviewer of IEEE TDSC, TIFS, TSC, and Neural Network!
  • [Aug. 2024] I present our work on algorithmic purpose limitation at USENIX Security’24 in Philadelphia!
  • [Jul. 2024] Our paper on unveiling IP vulnerabilities of GAN-based distributed ML is accepted by CIKM’24!
  • [Jun. 2024] I am the winner of the Richard Jago Memorial Prize!
  • [Feb. 2024] Our paper on algorithmic purpose limitation is accepted by USENIX Security’24!
  • [Aug. 2023] Our paper on formalizing neural network perturbation is accepted by ICFEM’23!
  • [Apr. 2023] I become a Ph.D. student at UQ supervised by A/Prof. Guangdong Bai and Dr. Jason Xue!
  • [Dec. 2022] I am successful in applying for CSIRO’s Data61 full scholarship with a top-up!
  • [Jan. 2021] Our paper on the LSTM-based hybrid water level prediction method is accepted by JASSE!

Selected Publications

Modifier Unlocked: Jailbreaking Text-to-Image Models Through Prompts
Shuofeng Liu, Mengyao Ma, Minhui Xue, Guangdong Bai. Oakland 2025
PDF BibTex

Being Transparent is Merely the Beginning: Enforcing Purpose Limitation with Polynomial Approximation
Shuofeng Liu, Zihan Wang, Minhui Xue, Long Wang, Yuanchao Zhang, Guangdong Bai. USENIX Security 2024
PDF BibTex

Unveiling Intellectual Property Vulnerabilities of GAN-Based Distributed Machine Learning through Model Extraction Attacks
Mengyao Ma, Shuofeng Liu, M.A.P. Chamikara, Mohan Baruwal Chhetri, Guangdong Bai. CIKM 2024
PDF BibTex

Formalizing Robustness Against Character-Level Perturbations for Neural Network Language Models
Zhongkui Ma, Xinguo Feng, Zihan Wang, Shuofeng Liu, Mengyao Ma, Hao Guan, Mark Huasong Meng. ICFEM 2023
PDF BibTex

LSTM Based Hybrid Method for Basin Water Level Prediction by Using Precipitation Data
Shuofeng Liu, Puwen Lei, Koji Koyamada. JASSE 2021
PDF BibTex