Guangsheng Yu (Saber)

Guangsheng Yu (Saber)

University of Technology Sydney

I am a Lecturer at University of Technology Sydney where I also obtained my PhD. Before that, I was a Postdoctoral Fellow at the CSIRO Data61 working with Dr. Xiwei Xu, Prof. Shiping Chen, and Dr. Wei Ni. I received my B.Eng. and M.Eng. degrees in Telecommunication Network Engineering from the University of New South Wales.

I have expertise in the reliability, trust, and scalability of data management systems, with a focus on security and privacy analysis in ML, including attacks and defenses, machine unlearning, incentivized/decentralized ML, and privacy-preserving blockchains. I have published 50+ papers at the top international journals and conferences with total .

If you are interested in my research and keen to collaborate or discuss potential PhD opportunities, please feel free to contact me via email. Much welcome!

🔥 News

  • 24.06.2025:  🎉🎉 Thrilled to announce that our paper, “Is Your AI Truly Yours? Leveraging Blockchain for Copyrights, Provenance, and Lineage”, has been accepted by IEEE Transactions on Service Computing (TSC)!
  • 16.06.2025:  🎉🎉 Thrilled to announce that our paper, BlockFUL: Enabling Unlearning in Blockchained Federated Learning, has been accepted by IEEE Transactions on Information Forensics & Security (TIFS)!
  • 04.06.2025:  🎉🎉 Thrilled to announce that our paper, “Understanding BRC-20: Hope or Hype”, has been accepted by IEEE Transactions on Computational Social Systems (TCSS)!
  • 25.05.2025:  🎉🎉 Thrilled to announce that our paper, Maximizing NFT Incentives: References Make You Rich, has been accepted by IEEE Transactions on Service Computing (TSC)!
  • 01.05.2025:  🎉🎉 Thrilled to announce that our paper, “RobustLight:Improving Robustness via Diffusion Reinforcement Learning for Traffic Signal Control”, has been accepted by ICML’25!
  • 29.03.2025:  🌟 We’re beyond excited! 🌟 Absolutely thrilled to share that our paper, Split Unlearning, has been directly accepted to ACM CCS’25 — one of the top-4 venues in cybersecurity! Huge thanks to everyone who supported this work — we’re so proud and can’t wait to share more soon! 🎉🔒💻
  • 17.02.2025:  🎉🎉 Thrilled to announce that our paper, Understanding DAOs: An Empirical Study on Governance Dynamics, has been published online in IEEE Transactions on Computational Social Systems (TCSS)!
  • 04.01.2025:  🎉🎉 Thrilled to announce that our paper, CAN-Trace Attack: Exploit CAN Messages to Uncover Driving Trajectories, has been accepted by IEEE Transactions on Intelligent Transportation Systems (TITS)! Congrats to my PhD student, Xiaojie Lin!
  • 29.12.2024:  🎉🎉 Thrilled to announce that our paper, Accountability and Reliability in 6G O-RAN: Who is Responsible When it Fails?, has been accepted by IEEE Wireless Communications (WSM)!

📝 Publications

👾 AI, AI-Cybersecurity, AI4Blockchain, Blockchain4AI

CCS'25
sym

Split Unlearning

Guangsheng Yu, Yanna Jiang, Qin Wang, Xu Wang, Baihe Ma, Caijun Sun, Wei Ni, Ren Ping Liu.

  • This work is the first to propose, implement, and evaluate a practical unlearning framework for split learning.
  • 🎉 This work has been directly accepted to ACM CCS’25.
TNNLS'23
sym

IronForge: An Open, Secure, Fair, Decentralized Federated Learning

Guangsheng Yu, Xu Wang, Caijun Sun, Qin Wang, Ping Yu, Wei Ni, Ren Ping Liu.

  • This work is an implementation of ERC-5521 in an AI context.
ICML'25
sym

RobustLight: Improving Data Quality via Diffusion Reinforcement Learning for Traffic Signal Control

Mingyuan Li, Jiahao Wang, Guangsheng Yu, et al.

  • This work makes novel use of diffusion models to address advesrail attack and missing data in real-world traffic signal control systems.
arXiv
sym

Fishers Harvest Parallel Unlearning in Inherited Model Networks

Xiao Liu, Mingyuan Li, Xu Wang, Guangsheng Yu, et al.

  • This work is the first-ever to achieve fully parallel unlearning across models that are interlinked.
  • 💪 This work has been submitted to TIFS.

💰 Blockchain and Web3

TEM'19
sym

Enabling Attribute Revocation for Fine-Grained Access Control in Blockchain-IoT Systems

Guangsheng Yu, Xuan Zha,Xu Wang, Wei Ni, Kan Yu, Ping Yu, J. Andrew Zhang, Ren Ping Liu, Y. Jay Guo.

TCSS'24
sym

Toward Web3 Applications: Easing the Access and Transition

Guangsheng Yu, Xu Wang, Qin Wang, Tingting Bi, Yifei Dong, Ren Ping Liu, Nektarios Georgalas, Andrew Reeves.

🚗 Applied Cybersecurity

📖 Standards

We are deeply involved in the Ethereum community, contributing, collaborating, and growing together with this amazing ecosystem! 🚀✨💎🙌 #Ethereum

💻 Teaching

  • Coord./Lect./Tut., 41900/42000 Cryptography, University of Technology Sydney. Large-scale course with over 450 students.
  • Coord./Lect./Tut., 41092 Network Fundamentals, University of Technology Sydney. Fundamentals course for all Cybersecurity students.
  • Coord./Lect., 65325 Digital Trace and Identity, University of Technology Sydney. Core course for Forensic students (Forensic-oriented).
  • Coord./Lect., 420108 Cybersecurity Management, University of Technology Sydney. OPM course for industry.
  • Tut., 43010 Cyber Threat Intelligence and Incident Response, University of Technology Sydney. SecOps course collaborated with ThreatDefense.
  • Tut., 48436 Digital Forensics, University of Technology Sydney. Advanced course for Forensic students (Cybersecurity-oriented).
  • Curric. Des., 43030 Professional Practice in Computing (Cybersecurity), University of Technology Sydney. Aimed for 500 students in two years.