The field of differentially private data generation and analysis is rapidly advancing, with a focus on developing innovative methods to ensure privacy while maintaining data utility. Recent research has explored the use of smooth sensitivity to improve the accuracy of differentially private selection mechanisms, as well as the impact of data domain extraction on synthetic data privacy. Additionally, there is a growing interest in hybrid modeling approaches that combine agent-based models with partial differential equation models to simulate the spatial spread of infectious diseases. Noteworthy papers in this area include:

• Differentially Private Selection using Smooth Sensitivity, which proposes the Smooth Noisy Max mechanism to achieve provably tighter expected errors.
• Understanding the Impact of Data Domain Extraction on Synthetic Data Privacy, which examines the role of data domain extraction in generative models and its impact on privacy attacks.
• A Hybrid ABM-PDE Framework for Real-World Infectious Disease Simulations, which presents a hybrid modeling approach to simulate the spatial spread of infectious diseases.
• WorldMove, a global open data for human mobility, which introduces a large-scale synthetic mobility dataset covering over 1,600 cities across 179 countries.
• Beyond the Generative Learning Trilemma, which explores the potential of Deep Generative Models in producing synthetic data that satisfies the Generative Learning Trilemma.
• Prototype-Guided Diffusion for Digital Pathology, which proposes a prototype-guided diffusion model to generate high-fidelity synthetic pathology data at scale.