Baicheng Chen / 陈柏成
About
Hi! I am Baicheng (Danny) Chen, a senior undergraduate student majoring in Computer Science and Engineering (CSE) at the Chinese University of Hong Kong, Shenzhen.
My research interests lie at the intersection of Trustworthy AI and Computer Security, focusing on the dual challenge of ensuring AI security while leveraging it for specialized security applications. Currently, I am working on Agentic AI security and the application of agent in cryptography.
I am fortunate to be advised by Prof. Baoyuan Wu at CUHK-Shenzhen. And I am also grateful to collaborate with Prof. Furong Huang (UMD), Prof. Tianxing He (IIIS, THU) and Prof. Tianhao Wang (UVA) along my research journey.
Currently, I am actively seeking for PhD positions (2027 Fall) and internship opportunities in the field of Trustworthy AI. If you are interested in my research and would like to chat with me further, please email me.
I am also open to any kind of collaborations and discussions, please feel free to reach out to me.
🔥 News
Research Interests
My research goal is to build AI systems that remain secure as they become increasingly autonomous in real world environments. I work at the intersection of AI and Security, studying how agentic deployment changes existing security assumptions, how to build safeguards that remain effective in practice, and how increasingly capable AI systems reshape both security risks and capabilities.
My recent research topics include:
- Security gaps in modern agentic AI deployments: [AdapAction] [ATU]
- Scalable safeguards and accountability mechanisms for AI systems: [VGMShield] [ATU]
- AI's broader security and societal impact: [CREBench] [VirtualCrime]
Selected Publications
Only some of my papers are listed. See more on the Papers page or Google Scholar.
(* indicates equal contribution)
Baicheng Chen*, Yu Wang*, Ziheng Zhou*, Xiangru Liu, Juanru Li, Yilei Chen, Tianxing He
CREBench: Evaluating Large Language Models in Cryptographic Binary Reverse Engineering
COLM 2026
A benchmark to evaluate LLMs' capabilities in cryptographic binary reverse engineering.
Baicheng Chen*, Mingda Zhang*, Min Zhang, Haizhou Li, Baoyuan Wu
AdapAction: Adaptive Target Action Backdoor Attack against GUI Agents
CVPR 2026
A novel backdoor attack against LLM-based GUI agents.
Experiences
The Chinese University of Hong Kong, Shenzhen
Undergraduate Student @ School of Data Science
Advised by Prof. Baoyuan Wu
GPA: 3.82/4.0
Tsinghua University
Research Intern @ IIIS (a.k.a. Yao Class) & Shanghai Qi Zhi Institute
Advised by Prof. Tianxing He and Prof. Yilei Chen
University of California, Berkeley
Visiting Student @ EECS
GPA: 4.0/4.0
Services
- Reviewer: COLM 2026, AAAI 2027, ARR August 2026, ICLR 2027
- Peer Advisor, School of Data Science, CUHK-Shenzhen (2024 – 2025, 2025 – 2026, 2026 – 2027)
Miscellaneous
- I am passionate about video editing and color grading 🎬. Selected work is available on Bilibili channels of Qingjiu TV Station and the CUHK-Shenzhen Student Union.