Hi! ๐ฅฐ Iโm Sen Nie (่ๆฃฎ), a first-year Masterโs student at the Visual Information Processing and Learning (VIPL) group, Institute of Computing Technology (ICT), Chinese Academy of Sciences (CAS). I also conduct research at the State Key Laboratory of AI Safety, under the supervision of Profs. Jie Zhang, Shiguang Shan (IEEE Fellow), and Xilin Chen (IEEE Fellow).
My research is currently focused on the frontier of AI Safety. Iโm passionate about building robust and trustworthy AI systems and am always open to collaboration or a good academic chat!
๐ฅ News
- Mar 2026 ๐ Our work BIMM has been accepted by TCSVT 2026. The paper will be available soon.
- Feb 2026 ๐ Paper V-Attack is accepted by CVPR 2026! Thanks to all co-authors!
- Jan 2026 ๐ฅ Our team AIST_ICT achieved 1st place in the Anti-BAD: An Anti-Backdoor Challenge for Post-Trained Large Language Models (IEEE SaTML 2026)! Congratulations to all team members!
- Jun 2025 ๐ Graduated from Wuhan University. Honored as an Outstanding Graduate and received the Outstanding Undergraduate Thesis award.
๐ Publications
Contrastive Spectral Rectification: Test-Time Defense towards Zero-shot Adversarial Robustness of CLIP
Sen Nie, Jie Zhang, Zhuo Wang, Shiguang Shan, Xilin Chen
arXiv preprint arXiv:2601.19210, 2026 (under review...)
CSR provides efficient, state-of-the-art adversarial robustness for VLMs by rectifying spectral inconsistencies during test time.
Sen Nie, Jie Zhang, Zhuo Wang, Shiguang Shan, Xilin Chen
arXiv preprint arXiv:2601.19210, 2026 (under review...)
CSR provides efficient, state-of-the-art adversarial robustness for VLMs by rectifying spectral inconsistencies during test time.
V-Attack: Targeting Disentangled Value Features for Controllable Adversarial Attacks on LVLMs
Sen Nie, Jie Zhang, Jianxin Yan, Shiguang Shan, Xilin Chen
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026
A novel adversarial attack method that targets the value features in transformer-based vision-language models to achieve precise local semantic manipulation, outperforming SOTA by 36% in success rate.
Sen Nie, Jie Zhang, Jianxin Yan, Shiguang Shan, Xilin Chen
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026
A novel adversarial attack method that targets the value features in transformer-based vision-language models to achieve precise local semantic manipulation, outperforming SOTA by 36% in success rate.
๐ Awards
- National First Prize (Top 0.5%) and National Demonstration Paper in CUMCM, for Automatic Pricing and Replenishment Decisions for Vegetable Products Based on Price Elasticity
- First Prize (2nd Place), Chinese Collegiate Computing Competition (Central South Region)
- First Prize, Huazhong Cup Mathematical Modeling Contest (HCMCM) (Top 0.4%)
- Scholarships: Lei Jun Computer Research Scholarship (Top 1%), NSFOCUS (LvMeng) Scholarship (Top 1%), First-Class Scholarship of Wuhan University (Top 5%)
๐ Honors
- Awarded the universityโs highest graduation distinction (Outstanding Graduate)
- Recognized for exemplary research (Outstanding Undergraduate Thesis)
- Honored as a Merit Studentโ for overall excellence (2022-2024)
๐ Education
- Sep. 2025 โ Present, M.S. Candidate, Institute of Computing Technology (ICT), Chinese Academy of Sciences (CAS), Beijing
- Sep. 2021 โ Jul. 2025, B.Eng. in Cyberspace Security, Wuhan University, Wuhan