–
Research Intern
Mentors: Xiaoting Qin and Fangkai Yang
Agent security, RSI attack
Northwestern University
Nan Yan
PhD student in Computer Science
- LLM security
- Agent security
- AI4Science
I am a first-year PhD student at Northwestern University, advised by Prof. Yan Chen. My current research focuses on the safety and security of recursive self-improvement (RSI) and AI agents. More recently, I have also begun exploring AI for Science (AI4Science).
Previously, I received my M.Eng. from Wuhan University, working with Prof. Yuqing Li and Prof. Jing Chen, and my B.Eng. from Shandong University.
Always happy to discuss research and collaborations.
Recently
News
Won Third Prize in the 2nd National Open Source Security Award Program.
Our paper was accepted to Findings of ACL 2026. Congratulations to Haoran!
Our paper was accepted to KDD 2026. Congratulations to my collaborators!
Our paper was accepted to ICML 2026. Congratulations to my collaborators!
Earlier updates 9
Received the National Scholarship.
Won Second Prize in the 1st National Open Source Security Award Program. Code.
Received a USENIX Security 2025 Student Grant. Thank you for the generous support!
Received a Cyberspace Security Innovation Grant. Thank you for the generous support!
Our paper was accepted to USENIX Security 2025. Congratulations to my collaborators!
Our paper was accepted to IEEE Transactions on Dependable and Secure Computing. Congratulations to my collaborators!
Received the National Scholarship.
Presented FedPHE at IEEE INFOCOM 2024.
Our paper was accepted to IEEE INFOCOM 2024. Congratulations to my collaborators!
Research
Publications
Selected work
-
ICML 2026
-
KDD 2026
-
USENIX Security 2025
LLM security
EmbedX: Embedding-Based Cross-Trigger Backdoor Attack Against Large Language Models
Investigating embedding-based backdoors that generalize across triggers in large language models.
More publications
-
Findings of ACL 2026
-
IEEE TDSC 2025
FedPHE: A Secure and Efficient Federated Learning via Packed Homomorphic Encryption
The first author is the advisor.
-
IEEE INFOCOM 2024
Efficient and Straggler-Resistant Homomorphic Encryption for Heterogeneous Federated Learning
Along the way
Experience
–
Research Intern
Working with Prof. Jiarong Xing
Agent security, memory attack
Recognition
Scholarships & honors
- Cyberspace Security Innovation GrantTOPSEC '23, Huawei '25 & DiDi '262023, 2025, 2026
- Outstanding Graduate AwardShandong University '23 & Wuhan University '262023, 2026
- National ScholarshipTop 0.2% nationwide · Ministry of Education, China2024, 2025
- BYD ScholarshipTop 3 in the Department of CSE · Wuhan University2025
- USENIX Security Student Grant2025
- Merit StudentTop 1% · Wuhan University2024, 2025
- First Class ScholarshipAward rate: 5% university-wide · Wuhan University2024, 2025
Earlier honors 2
- Alumni Council RepresentativeDepartment of CSE, 6/101 · Shandong University2023
- The Power of Role Models Academy Person of the YearTop 1 in the Department of CSE · Shandong University2023