Nan Yan bio photo

Email

LinkedIn

Github

Google Scholar

About Me

Here is Nan Yan (闫楠).

I am a first-year PhD student in the Department of Computer Science at Northwestern University, under the supervision of Prof. Yan Chen. Before that, I received my M.Eng. degree in Wuhan University, collaborating with Prof. Yuqing Li and Prof. Jing Chen, and received my B.Eng. degree from Shandong University.

My research interests lie in Trustworthy AI, with a particular focus on studying the security, safety, and privacy of machine learning systems. I work on developing innovative solutions for privacy-preserving federated learning, and I am especially interested in emerging challenges in the security of large language models (LLMs) and AI agents.

Outside the digital realm, I channel my energy into 🏊swimming and 💪fitness.

If you are interested in any aspect of me, I am always open to discussions and collaborations. Feel free to reach out to me at 📬nanyan [at] u [dot] northwestern [dot] edu

Research Interests

  • Security of generative AI and LLM agents: Exploring the security of large language models (LLMs), including backdoor attacks, jailbreak techniques, and emerging threats in LLM-powered agents.
  • Privacy: Investigating the use of Differential Privacy (DP) in the fine-tuning of LLMs to enhance privacy guarantees.
  • Federated Learning: Enhancing the efficiency and robustness of federated learning systems, with a focus on privacy-preserving techniques such as Homomorphic Encryption (HE) and DP.

News

Publications

  • Federated LoRA Fine-Tuning with Pipelined Error-Mitigated Aggregation and Matrix-Wise Freezing Haoran Wang, Xiong Wang, Yuqing Li, Jing Chen, Junyi Zhang, Nan Yan, Kun He, Wei Wang, in Findings of ACL, 2026. [pdf] [code]

  • Efficient and Differentially Private Federated LLM Fine-Tuning on Heterogeneous Clients

    Nan Yan, Yuqing Li, Jing Chen, Xiong Wang, Wei Wang, Kun He, Ruiying Du, Shuhua Li, in Proc. SIGKDD, 2026. [pdf] [code] [video]

  • MemIncept: Steering LLM Agents via Cooperative Stealthy Memory Injections

    Nan Yan, Qian Lou, Jiarong Xing, in Proc. ICML, 2026. [pdf] [code]

  • EmbedX: Embedding-Based Cross-Trigger Backdoor Attack Against Large Language Models

    Nan Yan, Yuqing Li, Xiong Wang, Jing Chen, Kun He, Bo Li, in Proc. USENIX Security, 2025. [pdf] [code] [video] [media1] [media2] [media3]

  • FedPHE: A Secure and Efficient Federated Learning via Packed Homomorphic Encryption

    Yuqing Li, Nan Yan, Jing Chen, Xiong Wang, Jianan Hong, Kun He, Wei Wang, Bo Li, in IEEE Transactions on Dependable and Secure Computing (TDSC), 2025. [pdf] [code] [html] [media] (The first author is the advisor)

  • Efficient and Straggler-Resistant Homomorphic Encryption for Heterogeneous Federated Learning

    Nan Yan, Yuqing Li, Jing Chen, Xiong Wang, Jiannan Hong, Kun He, and Wei Wang, in IEEE Conference on Computer Communications (INFOCOM), 2024. [pdf] [code] [slides] [html] [media]

Experiences

Microsoft:

Research Intern, Mentor: Xiaoting Qin and Fangkai Yang, April 2026 - September 2026

Project: Agent security in prompt injection

Rice University:

Research Intern, working with Prof. Jiarong Xing, June 2025 - February 2026

Project: Agent security in memory poisoning attack

Scholarships and Honors

  • Cyberspace Security Innovation Grant, TOPSEC’23 & Huawei’25 & DiDi’26 (2023, 2025, 206)
  • Outstanding Graduate Award, Shandong University’23 & Wuhan University’26 (2023, 2026)
  • BYD Scholarship (Top 3 in Dept.CSE), Wuhan University (2025)
  • USENIX Security’25 Student Grant (2025)
  • National Scholarship (Top 0.2% nationwide), Ministry of Education, China (2024, 2025)
  • Metrit Student (Ranking: Top 1%) , Wuhan University (2024, 2025)
  • First Class Scholarship (Award Rate: 5% school-wide), Wuhan University (2024, 2025)
  • Alumni Council Representative (Dept. CSE, 6/101), Shandong University (2023)
  • The Power of Role Models Academy Person of the Year (Dept. CSE, Top 1), Shandong University (2023)
  • Excellent Student Cadre, Shandong University (2022)
  • Second Class Scholarship (Award Rate: 10% school-wide), Shandong University (2020, 2022)
  • Merit-Based Scholarship, Shandong University (2020, 2022×3)

Competition Awards

  • Third Prize of the 2-nd National Open Source Security Award Program, Cyber Security Association, China (2026)
  • Second Prize of the 1-st National Open Source Security Award Program, Cyber Security Association, China (2025)
  • Second Prize of The 7-th National College Cryptography Mathematics Contest, Chinese Association for Cryptologic Research, China (2022)
  • Second Prize of The 15-th National College Student Information Security Competition, Cyber Security Association, China (2022)
  • First Prize of The 7-th National College Cryptography Mathematics Contest, Chinese Association for Cryptologic Research, North China Division (2022)
  • First Prize of The 31-th China Undergraduate Mathematical Contest in Modeling, Shandong Province (2021)

Patents and Software Copyrights

  • “Differential privacy-based heterogeneous federal fine tuning language model construction method and system”, China Patent Application ZL 2024 1 1379992.3, PatentGrant (Sep 2025)
  • ““Large language model training method and system based on elastic federated low-rank adaptive fine-tuning”, China Patent Application CN119443311A, PatentPending (Feb 2025)
  • “Method for constructing a vertical federated learning system based on participant selection and parameter freezing”, China Patent Application CN202411465068.7, PatentPending (Jan 2025)
  • “Cross-silo heterogeneous federated learning system based on homomorphic encryption V1.0”, China Software Copyrights 2024SR1516588 (Oct 2024)
  • “Method and device for constructing cross-silo heterogeneous federated learning system based on homomorphic encryption”, China Patent Application CN117892322A, PatentPending (Apr 2024)
  • “Network traffic obfuscation method, device, equipment and medium”, China Patent Application CN117749402A, PatentPending (Mar 2024)