Research Overview
Understanding of AI
We study the theoretical foundations of modern AI. Our goal is to understand how AI works.
Generative Modeling
Developing the fundamental algorithms of generative AI, such as diffusion models.
Generalization
Analyzing loss landscape geometry to understand when and why deep networks generalize.
Memorization
Investigating why memorization happens in AI and how it is related to generalization.
Privacy & Safety Issues of AI
We develop privacy- and safety-preserving machine learning algorithms. Our work spans formal privacy guarantees via differential privacy, machine unlearning for generative AI, and defenses against AI attacks.
Differential Privacy
Optimizing the utility-privacy tradeoff for ML models trained with DP-SGD, including private SVMs and synthetic data generation.
Machine Unlearning
Efficiently removing specific data or concepts from trained models (including diffusion models) without full retraining.
Attack Robustness
Understanding robustness and building resilience against attacks such as adversarial attacks and membership inference.
Industrial Applications
We apply trustworthy AI techniques to real-world industrial challenges.
Time Series
Robust forecasting and synthetic data generation for noisy, scarce sequential data.
Finance
Asset price modeling with generative diffusion models and applications of AI in financial fields.
Manufacturing
Generative AI for industrial domain and privacy concerns for industrial data.