Scholar
Zihan Wang
Google Scholar ID: 5tre4J4AAAAJ
The University of Queensland and CSIRO's Data61
Citations & Impact
Citations
511
H-index
6
i10-index
5
Publications
11
Co-authors
8
list available
Publications
Resume (English only)
Academic Achievements
- Publications:
- - WWW'25: AI Model Modulation with Logits Redistribution
- - OOPSLA'25: Convex Hull Approximation for Activation Functions
- - S&P'24: CoreLocker: Neuron-level Usage Control
- - TCSS: Data Hiding With Deep Learning: A Survey Unifying Digital Watermarking and Steganography
- Awards:
- - Google PhD Fellowship, 2024
- - Supported by RTP Scholarship and CSIRO’s Data61 Top-up Scholarship
Research Experience
- Serving as the HDR representative for the CSS discipline. Involved in multiple research projects including AI model usage control and neural network verification.
Education
- Graduated from the University of Adelaide with a B.CS (Adv.) in 2022. Currently a Ph.D. student at the University of Queensland, supervised by A/Prof. Guangdong Bai and Dr. Jason Xue.
Background
- Research interests: tackling real-world security & privacy issues of ML systems in a formally verifiable manner. Works closely with colleague Zhongkui Ma on neural network verification (NNV). Current projects include AI Model Usage Control, Neuron-level Usage Control (CoreLocker), Logit-level Model Modulation (AIM), Neural Network Verification (WRAACT), Algorithmic Purpose Limitation (AlgoSpec), Token-level Gradient Inversion (Grab), Multi-modal Membership Inference (M⁴I), Model Pruning (CoreLocker), and Knowledge Distillation (BPKD). Currently serving as the HDR representative for the CSS discipline.
Miscellany
- Personal interests not mentioned
