Privacy Utility

Privacy utility research focuses on balancing the protection of sensitive data with the usefulness of that data for machine learning tasks. Current efforts concentrate on developing differentially private algorithms and mechanisms, including those tailored for specific model architectures like graph neural networks and boosted decision trees, as well as exploring synthetic data generation and federated learning approaches. This field is crucial for responsible data usage, enabling the development of powerful machine learning models while mitigating privacy risks in various applications, from healthcare to smart grids.