Short Biography

Zhuqing (Zuki) Liu is an Assistant Professor in the Department of Computer Science and Engineering at the University of North Texas (UNT). She received her Ph.D. degree from the Department of Electrical and Computer Engineering at The Ohio State University in 2024, advised by Prof. Jia (Kevin) Liu.

Immediate Openings for Ph.D. Students: I am recruiting Ph.D. students to start in Fall 2027 or Spring 2028. If the topics on my Research page interest you, email me a single message with your CV, transcripts, and any papers you have written; GRE and TOEFL scores are welcome but not required at this stage. A short paragraph on which of my papers you found interesting, and why, tells me far more than a generic cover letter. Volume makes individual replies impossible, so silence is not a rejection. Admission decisions are made by the department, so please also apply through the UNT CSE graduate program.

Research Interests

  • AI Security and Large Language Models: poisoning and backdoor attacks, responsibility attribution, secure retrieval-augmented generation, LLM agent forensics
  • First-Order Stochastic Optimization for Machine Learning: bilevel, min-max, and multi-objective learning with provable convergence and sample complexity
  • Communication Efficiency in Distributed/Federated/Decentralized Learning: Byzantine robustness, differential privacy, split and asynchronous federated learning

News

  • A paper on post-hoc patching of backdoored large language models has been accepted by USENIX Security Symposium 2026.
  • Two papers, on forensic attribution of backdoored code completions and on failure localization in LLM-based multi-agent systems, have been accepted by COLM 2026 (acceptance rate: 29%).
  • I received a $400,000 grant from the State of Texas as the Lead PI (09/2026 – 07/2028), in collaboration with Kai Zhang and Teresa Wagner.
  • Our project "Towards Secure Retrieval-Augmented Generation" has been awarded 7,000 NVIDIA H200 GPU-hours on TACC Vista by the NAIRR Pilot (NAIRR250513). I am honored to serve as the Lead PI, in collaboration with Minghong Fang.
  • I received an allocation of 20,000 SUs on TACC Lonestar6 for the project "Complex-Structured Unlearning in Machine Learning" as the Sole PI (08/2025 – 08/2026).
  • A paper on mitigating backdoor attacks in split learning has been accepted by The Web Conference (WWW) 2026 (acceptance rate: 20.1%).
  • A paper on responsibility attribution for poisoned knowledge in retrieval-augmented generation has been accepted by IEEE Symposium on Security and Privacy 2026 (a "Security Big Four," acceptance rate: 13%).
  • I am honored to serve as a TPC member of ACM MobiHoc 2026.
  • I received $10,000 in CENG–COS Joint Proposal Development Seed Funding at UNT as the Lead PI, in collaboration with Yanxiao Han.
  • A paper on multi-objective bilevel learning has been accepted by AAAI 2026 (acceptance rate: 17.6%).
  • A paper on secure asynchronous federated learning has been accepted by ACM AsiaCCS 2026.
  • I was an invited speaker at the STEAM Learning Club in Texas outreach program, January 2026.
  • A paper on differentially private decentralized min-max optimization has been accepted by ACM MobiHoc 2025 (acceptance rate: 23%).
  • A paper on fast convergence and low sample complexity in stochastic multi-objective learning has been accepted by UAI 2025 (acceptance rate: 31%).
  • I was an invited speaker at the UNT Data Science Series, March 2025.
  • A paper on decentralized bilevel optimization without lower-level strong convexity has been accepted by ICLR 2025 (acceptance rate: 32%).
  • A paper on Byzantine-robust aggregation rules has been accepted by NDSS 2025 (a "Security Big Four," acceptance rate: 15%).
  • Three papers have been accepted by The Web Conference (WWW) 2025 (acceptance rate: 19.8%).
  • I joined the Department of Computer Science and Engineering at the University of North Texas as a tenure-track Assistant Professor.
Back to Top ↑