Private Probabilistic

Private probabilistic methods aim to develop machine learning models that both protect sensitive data and produce accurate results. Current research focuses on integrating differential privacy mechanisms into various learning paradigms, including federated learning and generative models, often employing techniques like Gaussian mechanisms and gradient perturbation to achieve privacy guarantees while maintaining model utility. This work is crucial for enabling responsible data analysis and model training in sensitive domains like healthcare and finance, where privacy is paramount. The development of robust and efficient private probabilistic models is driving advancements in both theoretical understanding of privacy and practical applications of machine learning.

Papers

January 18, 2024

Differentially Private and Adversarially Robust Machine Learning: An Empirical Evaluation
Janvi Thakkar, Giulio Zizzo, Sergio Maffeis
Adversarial Training Membership Inference Attack Empirical Evaluation Differentially Private Robust Adversarial Learning Simultaneous Attack Private Probabilistic

May 23, 2023

Fair Differentially Private Federated Learning Framework
Ayush K. Varshney, Sonakshi Garg, Arka Ghosh, Sargam Gupta
Differential Privacy Homomorphic Encryption Private Federated Learning Private Probabilistic

May 18, 2023

Private Gradient Estimation is Useful for Generative Modeling
Bochao Liu, Pengju Wang, Weijia Guo, Yong Li, Liansheng Zhuang, Weiping Wang, Shiming Ge
LeArning Abstract Differential Privacy Image Generation Privacy Preserving Image Private Generative Model Private Probabilistic

April 15, 2022

A Differentially Private Probabilistic Framework for Modeling the Variability Across Federated Datasets of Heterogeneous Multi-View Observations
Irene Balelli, Santiago Silva, Marco Lorenzi
Multi Center Study Hierarchical Latent Federated Data Multiview Observation Private Probabilistic

Private Probabilistic

Papers

Differentially Private and Adversarially Robust Machine Learning: An Empirical Evaluation

Fair Differentially Private Federated Learning Framework

Private Gradient Estimation is Useful for Generative Modeling

A Differentially Private Probabilistic Framework for Modeling the Variability Across Federated Datasets of Heterogeneous Multi-View Observations