Biography
Hi, I'm Xin Wang, a Ph.D. Candidate of FVL Lab in the School of Computer Science at Fudan University, supervised by Prof. Xingjun Ma and Prof. Yu-Gang Jiang. Recently, I am broadly interested in safety and privacy aspects of machine learning with a recent focus on large language models. Most of my past works are in the domain of trustworthy machine learning, particularly adversarial examples and robustness of machine learning algorithms.
Feel free to reach me at xinwang22 [at] m [dot] fudan [dot] edu [dot] cn, if you are interested in potential collaborations.
Publication
- TAPT: Test-Time Adversarial Prompt Tuning for Robust Inference in Vision-Language Models
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, USA, 2025
- AdvQDet: Detecting Query-Based Adversarial Attacks with Adversarial Contrastive Prompt Tuning
ACM Multimedia (MM), Melbourne, Australia, 2024
- FreezeVLA: Action-Freezing Attacks against Vision-Language-Action Models
arXiv, 2025
- SafeWork-R1: Coevolving Safety and Intelligence under the AI-45° Law
Technical Report, 2025
- Adversarial Prompt Tuning for Vision-Language Models
European Conference on Computer Vision (ECCV), Milano, Italy, 2024
- SafeVid: Toward Safety-Aligned Video Large Multimodal Models
Advances in Neural Information Processing Systems (NeurIPS), San Diego, USA, 2025
- Argus Inspection: Do Multimodal Large Language Models Possess the Eye of Panoptes?
ACM Multimedia (MM), Dublin, Ireland, 2025
- Safety at Scale: A Comprehensive Survey of Large Model Safety
Foundations and Trendss® in Privacy and Security, 2025
- Lossless Medical Image Compression Based on Anatomical Information and Deep Neural Networks
Biomedical Signal Processing and Control, 2022
- Web-Based Technology for Remote Viewing of Radiological Images: App Validation
Journal of Medical Internet Research, 2020
- Evolve the Method, Not the Prompts: Evolutionary Synthesis of Jailbreak Attacks on LLMs
arXiv, 2025
- Simulated Ensemble Attack: Transferable Jailbreaks Across Fine-tuned Vision-Language Models
arXiv, 2025
- NAP-Tuning: Neural Augmented Prompt Tuning for Adversarially Robust Vision-Language Models
arXiv, 2025
- SafeEvalAgent: Toward Agentic and Self-Evolving Safety Evaluation of LLMs
arXiv, 2025
- DarkLLaVA: Scalable Adversarial Attack with Large Language Models
arXiv, 2025
- Adversarial Prompt Distillation for Vision-Language Models
arXiv, 2024
- A2RM: Adversarial-Augmented Reward Model
arXiv, 2025
- LeakyCLIP: Extracting Training Data from CLIP
arXiv, 2025
- Imperceptible Jailbreaking against Large Language Models
arXiv, 2025
- BackdoorVLM: A Benchmark for Backdoor Attacks on Vision-Language Models
arXiv, 2025
- DAVID-XR1: Detecting AI-Generated Videos with Explainable Reasoning
arXiv, 2025
Hornors & Awards
- First Class Award Scholarship of Fudan University, 2025
- Outstanding Student of Fudan University, 2023
- iDASH Privacy & Security Challenge Track 3: Confidential Computing, Ranked 1st, 2021
- Outstanding Graduate Award of Central China Normal University, 2021
- Outstanding Master's Thesis Award of Central China Normal University, 2021
Professional Service
- Conference Reviewer for ICLR, NeurIPS, CVPR, ICCV, AAAI, ACM MM, et al.
- Journal Reviewer for IJCV, TIP, TCSVT, TNNLS, et al.