Privacy & Security Lab
The Privacy & Security Lab focuses on privacy-preserving and secure data analysis for modern artificial intelligence and machine learning systems. Our goal is to develop practical methods with formal privacy and security guarantees while maintaining high data utility.
Research Areas
Differential Privacy (DP)
We design and analyze privacy-preserving mechanisms with rigorous theoretical guarantees, covering centralized, local (LDP), and shuffle models of differential privacy.
Privacy-Preserving Data Synthesis
We study differentially private data generation techniques for tabular, image, and time-series data using DP-GAN, DP-VAE, and DP-SGD to enable safe data sharing and reuse.
Privacy Risk and Security Analysis
We investigate privacy and security threats such as membership inference attacks, re-identification risks, and data poisoning attacks, and develop corresponding defense mechanisms.
Trustworthy and Secure AI Systems
We explore privacy-enhancing technologies in real-world applications including healthcare, smart cities, and data-sharing platforms to build trustworthy and compliant AI systems.