We are a research lab at Queen's University, working at the intersection of machine learning, cybersecurity, and privacy. Our work spans both attacking and defending intelligent systems.
- AI Security — Robustness of federated and split learning against poisoning, backdoor, and adversarial attacks
- Secure Generative AI — Preventing unauthorized manipulation of AI-generated images; copyright and ethical safeguards for diffusion models
- Image Watermarking — Invisible and robust watermarking for ownership verification and content authentication; watermark-preserving image editing
- Agentic AI — Security and privacy challenges in autonomous AI agent systems
| Paper | Venue | Code |
|---|---|---|
| MarkNull: Model-Agnostic Watermark Removal in AI-Generated Images via On-Manifold Latent Manipulation | USENIX Security 2026 | repo |
| SecureT2I: No More Unauthorized Manipulation on AI Generated Images from Prompts | ESORICS 2025 | repo |
| Are Watermarked Images Editable? SafeMark for Watermark-Preserving Text-Guided Image Editing | Preprint | repo |
| Evaluating Security and Robustness for Split Federated Learning against Poisoning Attacks | IEEE T-IFS 2024 | repo |
For inquiries, visit our lab website.
