[10/2024] 📑 One paper about Backdoor in Federated Learning is released.
[03/2024] 🔬 I will be an intern of ALPS in SBU this summer!
[12/2023] 📑 One paper about Generated Image Detection is released on the arXiv.

I am committed to building trustworthy AI systems capable of handling complex tasks in critical scenarios. I believe it is essential to address three key factors: robustness, safety, and interpretability. Driven by these considerations, I aim to explore the following research questions:
• How can we ensure that models maintain sufficient robustness during interactions with external environments?
• How can we minimize bias and harmful content in model outputs to promote fairness and reliability?
• How can we establish comprehensive benchmarks that measures the trustworthiness LLMs from multiple dimensions?

In my past years of research experience, I have also explored these research directions:
AIGC Detection: How to quickly and accurately distinguish between real content and AI-generated content to protect property and personal safety from the harm of false information.
Backdoor Attack: How to design new backdoor attack paradigms to expose security vulnerabilities in machine learning models, providing new perspectives for secure and trustworthy defense mechanisms.
LLM Security: How to enhance the robustness of large language models to ensure their outputs are free from bias and harmful content, enabling their application in safety-critical complex tasks.

⚾️ I am a big baseball fan of the Los Angeles Dodgers.
🎮 I am a big fan of Souls-like games, with Dark Souls 3 being my favorite.
🥇 I participated in physics and math olympiads in high school.
🏃‍♂️ I am a long-distance running enthusiast.
📷 I am a photography enthusiast, using a Fuji XT-5.