Yixin Wu

Im Oberen Werk 1
66386 St. Ingbert (Germany)
I’m a Ph.D. student at CISPA Helmholtz Center for Information Security, where I am fortunate to be advised by Prof. Michael Backes and Dr. Yang Zhang. Prior to coming to CISPA, I received my Bachelor’s degree from Sichuan University, where I daily worked with Prof. Cheng Huang. During my undergraduate, I was also a security engineer intern at Alibaba.
My research focuses on designing and developing trustworthy AI systems, ensuring they are safe, privacy-preserving, and secure. I am also interested in the responsible use of AI, with a focus on transparency and preventing the misuse of AI-generated content. Currently, I am passionate about building generative agents for security and privacy tasks, as well as social behavior simulation.
Research Interests
- Trustworthy AI
- Privacy:[USENIX'24],[PETS'24]
- Safety:[CCS'24],[USENIX'25b]
- Responsible use of AI
- Transparency in synthetic data usage:[USENIX'25a]
- Misuse of AI-generated content:[USENIX'25c]
- In-the-wild prompt analysis:[EMNLP'24]
- Generative Agents for Security, Privacy, and Social Behavior Simulation
Honors and Awards
- 2025ML and Systems Rising Star, MLCommons
- 2021Outstanding Graduate Honor, Sichuan University
- 2019National Scholarship, Ministry of Education of China
News
Mar 2025 | I was selected as a ML and Systems Rising Star! |
---|---|
Jan 2025 | Our paper titled “Synthetic Artifact Auditing: Tracing LLM-Generated Synthetic Data Usage in Downstream Applications” was accepted by Usenix Security 2025. See the website for more details! |
Jan 2025 | Our paper titled “On the Proactive Generation of Unsafe Images From Text-To-Image Models Using Benign Prompts” was accepted by Usenix Security 2025! |
Jan 2025 | Our paper titled “HateBench: Benchmarking Hate Speech Detectors on LLM-Generated Content and Hate Campaigns” was accepted by Usenix Security 2025! See the website for more details! |
Dec 2024 | I will join the PC of CCS 2025! |
Sep 2024 | Our paper titled “The Death and Life of Great Prompts: Analyzing the Evolution of LLM Prompts from the Structural Perspective” was accepted by EMNLP 2024! |
Aug 2024 | Our paper titled “Image-Perfect Imperfections: Safety, Bias, and Authenticity in the Shadow of Text-To-Image Model Evolution” was accepted by CCS 2024! |
May 2024 | Our paper titled “Link Stealing Attacks Against Inductive Graph Neural Networks” was accepted by PETS 2024! |
Sep 2023 | Our paper titled “Quantifying Privacy Risks of Prompts in Visual Prompt Learning” was accepted by Usenix Security 2024! |
Selected Publications
- Usenix SecurityIn USENIX Security Symposium (USENIX Security), 2025
- Usenix SecurityIn USENIX Security Symposium (USENIX Security), 2025
- CCSIn ACM Conference on Computer and Communications Security (CCS), 2024
- EMNLPIn Empirical Methods in Natural Language Processing (EMNLP), 2024