Microsoft — Research Intern
- Differentially private data synthesis for structured datasets with large language models.
- Hosts: Zinan Lin, Pei Zhou.
I am a fifth-year Ph.D. candidate in ECE at Carnegie Mellon University, advised by Giulia Fanti. I work on the theoretical foundations of machine learning, and the applications in large language models, synthetic data, AI safety, security and privacy, and federated learning.
Research Interests
Large Language Model · Synthetic Data · AI Safety · ML Privacy & Security · Model Evaluation · Federated Learning
Education
- Carnegie Mellon University — Ph.D., Electrical & Computer Engineering (2021 – present). Advisor: Giulia Fanti.
- Shanghai Jiao Tong University — B.S., Computer Science & Zhiyuan Honors Program (2016 – 2020)
Industry Experience
Amazon — Research Intern
- Adversarial attacks on vision–language models; evaluated stealth and efficacy.
- Host: Alireza Mehrtash.
JPMorgan Chase — AI Research Intern
- Privacy protection across multiple summary statistical properties for data sharing.
- Host: Mohsen Ghassemi.
Publications & Manuscripts
-
Struct‑Bench: A Benchmark for Differentially Private Structured Text Generation
-
Inferentially‑Private Private Information
-
MASAN: Enhancing Attack Stealth and Efficacy on Vision‑Language Models via Smart Noise
-
Linear Combination of Saved Checkpoints Makes Consistency and Diffusion Models Better
-
Evaluating Selective Encryption Against Gradient Inversion Attacks
-
Statistic Maximal Leakage
-
Guarding Multiple Secrets: Enhanced Summary Statistic Privacy for Data Sharing
-
Data Distribution Valuation
-
Mixture‑of‑Linear‑Experts for Long‑term Time Series Forecasting
-
Summary Statistic Privacy in Data Sharing
-
Towards a Defense against Backdoor Attacks in Continual Federated Learning
-
Efficient Distributed Steiner Tree Construction in Wireless Sensor Networks with Unreliable Links
-
De‑anonymizability of Social Network: Through the Lens of Symmetry
-
De‑anonymizing Social Networks with Overlapping Community Structure
-
Collective Influence Maximization in Mobile Social Networks
Honors & Awards
- Carnegie Institute of Technology Dean’s Fellow, Carnegie Mellon University, 2021
- Zhiyuan Distinguished Scholarship, Shanghai Jiao Tong University, 2020
- Zhiyuan College Honors Scholarship, SJTU, 2017–2019
- Academic Excellence Scholarship, SJTU, 2017–2019
- First prize, Chinese University Student Computer Design Competition, 2019