Xuankun Rong (容旋坤)

Short Bio

My name is Xuankun Rong (容旋坤), I am currently a first-year Ph.D. student at the School of Computer Science, Wuhan University, where I am fortunate to be advised by Prof. Mang Ye. I received my Bachelor's degree at the School of Cyber Science and Engineering, Wuhan University.
My research focuses on AI Safety.

News

2025.11: Two papers have been accepeted by AAAI 2026 (One Oral). See you in Singapore! 🇸🇬
2025.10: I am invited as a reviewer for CVPR 2026.
2025.09: One paper (BYE) has been accepeted by NeurIPS 2025. See you in San Diego! 🇺🇸
2025.05: Released our new work BYE on backdoor defense for MLLMs on arxiv — welcome to follow and share! ❤️
2025.05: One paper (CAN) has been accepeted by ICML 2025. 💪
2024.09: Won Nomination Award of College Students' Innovation and Entrepreneurship works in 2024 China Network Security Innovation and Entrepreneurship Competition.
2024.08: Won First Prize in the 17th National College Student Information Security Contest, along with the Most Innovative and Entrepreneurial Award (Top 0.1% nation-wide). 👏
2024.07: We advanced to the National Finals of the 17th National College Student Information Security Contest and are looking forward to achieving excellent results. 💪
2024.05: Won First Prize in the 17th China Undergraduate Computer Design Competition, Middle South Division (Top 3% division-wide).

Selected Publications

Survey

A Survey of Safety on Large Vision-Language Models: Attacks, Defenses and Evaluations
Mang Ye*, Xuankun Rong, Wenke Huang, Bo Du, Nenghai Yu, Dacheng Tao
arXiv, 2025
[Paper][Code]

Keeping Yourself is Important in Downstream Tuning Multimodal Large Language Model
Wenke Huang, Jian Liang, Xianda Guo, Yiyang Fang, Guancheng Wan, Xuankun Rong, Chi Wen,
Zekun Shi, Qingyun Li, Didi Zhu, Yanbiao Ma, Ke Liang, Bin Yang, He Li, Jiawei Shao, Mang Ye, Bo Du
arXiv, 2025
[Paper][Code]

2025

SafeGRPO: Self-Rewarded Multimodal Safety Alignment via Rule-Governed Policy Optimization
Xuankun Rong, Wenke Huang, Tingfeng Wang, Daiguo Zhou, Bo Du, Mang Ye*
arXiv, 2025
[Paper][Code]

Believe Your Eyes: Backdoor Cleaning without External Guidance in MLLM Fine-tuning
Xuankun Rong^†, Wenke Huang^†, Jian Liang, Jinhe Bi, Xun Xiao, Yiming Li, Bo Du, Mang Ye*
Conference on Neural Information Processing System (NeurIPS), CCF-A, 2025
[Paper][Code]

Probing Semantic Insensitivity for Inference-Time Backdoor Defense in Multimodal Large Language Model
Xuankun Rong, Wenke Huang, Wenzheng Jiang, Yiming Li, Wenxuan Wang, Mang Ye*
Association for the Advancement of Artificial Intelligence (AAAI), CCF-A, 2026
[Paper][Code]

PurMM: Attention-Guided Test-Time Backdoor Purification in Multimodal Large Language Models
Wenzheng Jiang^†, Ke Liang^†, Xuankun Rong^†, Jingxuan Zhou, Zhengyi Zhong, Guancheng Wan, Ji Wang
Association for the Advancement of Artificial Intelligence (AAAI), Oral, CCF-A, 2026
[Paper][Code]

CAN: Leveraging Clients As Navigators for Generative Replay in Federated Continual Learning
Xuankun Rong^†, Jianshu Zhang^†, Kun He, Mang Ye*
International Conference on Machine Learning (ICML), CCF-A, 2025
[Paper][Code]

Selected Awards and Honors

Elite Ph.D. Talent Program in Mathematics and Artificial Intelligence of Wuhan University, 2025
(武汉大学数智学科博士生拔尖人才培养计划, Award Rate: 7 Members School-wide)

Lei Jun Computer Graduate Scholarship, 2025
(雷军计算机研究生奖学金)

First Prize in the 17th National College Student Information Security Contest, 2024
(第十七届信息安全大赛全国一等奖)