👤 About Me

I’m Xinwei Liu, a Ph.D. candidate at the Institute of Information Engineering, Chinese Academy of Sciences (IIE, CAS), where I focus on advancing the frontiers of AI security and privacy protection. I am fortunate to be supervised by Prof. Xiaochun Cao (Dean of the School of Cyber Science and Technology at Sun Yat-sen University) and Prof. Hua Zhang.

I received my Bachelor’s degree from the School of Mathematics and Computer Science, Nanchang University in 2020, advised by Prof. Yuchao Tang. During my research journey, I also gained valuable industry experience as a research intern at Ant Group (2022-2023).

Research Focus: My work centers on developing trustworthy AI systems through multimodal data protection and proactive defense mechanisms. I specialize in:

• 🛡️ AI Security: Adversarial attacks, backdoor defenses, and poisoning strategies
• 🔒 Privacy Protection: Machine unlearning and data privacy for multimodal systems
• 🚨 Large Model Safety: Jailbreak attacks and defenses for LLM/VLM models

🔥 News

• 🎉 Nov 2025: Two papers are accepted in AAAI 2026!
• 🎉 Oct 2025: One paper on Security of VLM is accepted in T-IFS 2025!
• 🎉 Sep 2024: One paper on Privacy of Multi-modal Data accepted by ACM MM 2024!
• 🎉 Feb 2024: One paper on Backdoor Attack is accepted by T-IFS 2024!

📝 Publications

🏆 Selected Publications

AAAI 2026

GeoShield: Safeguarding Geolocation Privacy from Vision-Language Models via Adversarial Perturbations

Xinwei Liu, Xiaojun Jia, Yuan Xun, Simeng Qin, Xiaochun Cao

PDF AAAI 2026

ACM MM 2024

Multimodal unlearnable examples: Protecting data against multimodal contrastive learning

Xinwei Liu, Xiaojun Jia, Yuan Xun, Siyuan Liang, Xiaochun Cao

PDF Code Proceedings of the 32nd ACM International Conference on Multimedia 2024 (ACM MM, 2024)

AAAI 2024

Does few-shot learning suffer from backdoor attacks?

Xinwei Liu, Xiaojun Jia, Jindong Gu, Yuan Xun, Siyuan Liang, Xiaochun Cao

PDF Code Thirty-Eighth AAAI Conference on Artificial Intelligence (AAAI 2024)

ECCV 2022

Watermark vaccine: Adversarial attacks to prevent watermark removal

Xinwei Liu, Jian Liu, Yang Bai, Jindong Gu, Tao Chen, Xiaojun Jia, Xiaochun Cao

PDF Code European Conference on Computer Vision 2022 (ECCV 2022)

AAAI 2026

The Emotional Baby Is Truly Deadly: Does your Multimodal Large Reasoning Model Have Emotional Flattery towards Humans?

Yuan Xun, Xiaojun Jia, Xinwei Liu, Hua Zhang

PDF AAAI 2026

TIFS 2025

CleanerCLIP: Fine-grained Counterfactual Semantic Augmentation for Backdoor Defense in Contrastive Learning

Yuan Xun, Siyuan Liang, Xiaojun Jia, Xinwei Liu, Xiaochun Cao

PDF IEEE Transactions on Information Forensics and Security (TIFS)

TIFS 2024

Minimalism is king! high-frequency energy-based screening for data-efficient backdoor attacks

Yuan Xun, Xiaojun Jia, Jindong Gu, Xinwei Liu, Qing Guo, Xiaochun Cao

PDF IEEE Transactions on Information Forensics and Security (TIFS 2024)

TMLR 2024

A Survey on Transferability of Adversarial Examples Across Deep Neural Networks

Jindong Gu, Xiaojun Jia, Pau de Jorge, Wenqain Yu, Xinwei Liu, Avery Ma, Yuan Xun, Anjun Hu, Ashkan Khakzar, Zhijiang Li, Xiaochun Cao, Philip Torr

PDF Transactions on Machine Learning Research 2024 (TMLR 2024)

CVPR Workshop

Universal watermark vaccine: Universal adversarial perturbations for watermark protection

Jianbo Chen, Xinwei Liu, Siyuan Liang, Xiaojun Jia, Yuan Xun

PDF Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshop (CVPR Workshop 2023)

JEI

Primal-dual algorithm to solve the constrained second-order total generalized variational model for image denoising

Xinwei Liu, Yuchao Tang, Yixuan Yang

PDF Journal of Electronic Imaging

📖 Preprints

See full list on Google Scholar

arXiv

PersGuard: Preventing Malicious Personalization via Backdoor Attacks on Pre-trained Text-to-Image Diffusion Models

Xinwei Liu, Xiaojun Jia, Yuan Xun, Hua Zhang, Xiaochun Cao

PDF arXiv 2025

🎖 Honors and Awards

• National Scholarship (Doctoral Level), Ministry of Education of China, 2024
• National Scholarship (Undergraduate Level), Ministry of Education of China, 2019
• Third Prize (Ranked 1st) in the 2025 Qiyuan Large Model Adversarial Challenge

🚩 Service

• Conference Reviewer: CVPR, NeurIPS, ICLR, ICCV, ECCV, ACM MM, AAAI
• Journal Reviewer: IEEE T-PAMI, IEEE TIFS, IEEE TIP, IEEE TDSC, Pattern Recognition

💻 Internships

• 2022.03 - 2023.06, Research Intern, Ant Group, China