Differential Privacy

Differential privacy (DP) is a rigorous framework for ensuring data privacy in machine learning by adding carefully calibrated noise to model training processes. Current research focuses on improving the accuracy of DP models, particularly for large-scale training, through techniques like adaptive noise allocation, Kalman filtering for noise reduction, and novel gradient processing methods. This active area of research is crucial for enabling the responsible use of sensitive data in various applications, ranging from healthcare and finance to natural language processing and smart grids, while maintaining strong privacy guarantees.

Papers

December 12, 2023

Communication Cost Reduction for Subgraph Counting under Local Differential Privacy via Hash Functions
Quentin Hillebrand, Vorapong Suppakitpaisarn, Tetsuo Shibuya
Differential Privacy Subgraph Structure Local Differential Privacy Large Graph Communication Cost Hash Function Graph Structure Information Subgraph Counting

December 11, 2023

Classification with Partially Private Features
Zeyu Shen, Anilesh Krishnaswamy, Janardhan Kulkarni, Kamesh Munagala
Classification Code Differential Privacy Private Data Privacy Guarantee Private Linear Regression Private Feature

December 10, 2023

QMGeo: Differentially Private Federated Learning via Stochastic Quantization with Mixed Truncated Geometric Distribution
Zixi Wang, M. Cenk Gursoy
Differential Privacy Privacy Guarantee Private Federated Learning Product Quantization Stochastic Quantization

December 9, 2023

Optimal Unbiased Randomizers for Regression with Label Differential Privacy
Ashwinkumar Badanidiyuru, Badih Ghazi, Pritish Kamath, Ravi Kumar, Ethan Leeman, Pasin Manurangsi, Avinash V Varadarajan, Chiyuan Zhang
Differential Privacy Novel Regression Privacy Utility Random Label

December 8, 2023

On the Inadequacy of Similarity-based Privacy Metrics: Reconstruction Attacks against "Truly Anonymous Synthetic Data''
Georgi Ganev, Emiliano De Cristofaro
Synthetic Data Differential Privacy PRIvacy Leakage Empirical Privacy Reconstruction Attack

December 5, 2023

December 2, 2023

PAC Privacy Preserving Diffusion Models
Qipan Xu, Youlong Ding, Xinxi Zhang, Jie Gao, Hao Wang
Diffusion Model Differential Privacy Privacy Protection Private Generative Model

November 30, 2023

Navigating Privacy and Copyright Challenges Across the Data Lifecycle of Generative AI
Dawen Zhang, Boming Xia, Yue Liu, Xiwei Xu, Thong Hoang, Zhenchang Xing, Mark Staples, Qinghua Lu, Liming Zhu
Artificial Intelligence Differential Privacy Generative AI Data Privacy Data Management Visual Foresight Protecting Privacy

November 27, 2023

Using Decentralized Aggregation for Federated Learning with Differential Privacy
Hadeel Abd El-Kareem, Abd El-Moaty Saleh, Ana Fernández-Vilas, Manuel Fernández-Veiga, asser El-Sonbaty
Differential Privacy Membership Inference Attack Personal Information Privacy Breach Decentralized Aggregation

November 24, 2023

Differentially Private SGD Without Clipping Bias: An Error-Feedback Approach
Xinwei Zhang, Zhiqi Bu, Zhiwei Steven Wu, Mingyi Hong
Differential Privacy Absolute Stance Bias Error Feedback Privacy Guarantee Private Stochastic Gradient Descent Gradient Clipping

November 23, 2023

DPSUR: Accelerating Differentially Private Stochastic Gradient Descent Using Selective Update and Release
Jie Fu, Qingqing Ye, Haibo Hu, Zhili Chen, Lulu Wang, Kuncan Wang, Xun Ran
Differential Privacy Stochastic Gradient Descent Private Stochastic Gradient Descent Privacy Preserving Machine Learning Gradient Update Private Model Training Model Update

November 22, 2023

November 20, 2023

November 15, 2023

Are Normalizing Flows the Key to Unlocking the Exponential Mechanism?
Robert A. Bridges, Vandy J. Tombs, Christopher B. Stanley
Differential Privacy Empirical Privacy Exponential Mechanism Private Machine Learning Private Optimization

November 14, 2023

Sparsity-Preserving Differentially Private Training of Large Embedding Models
Badih Ghazi, Yangsibo Huang, Pritish Kamath, Ravi Kumar, Pasin Manurangsi, Amer Sinha, Chiyuan Zhang
Full Model Differential Privacy Stochastic Gradient Descent Private Model Training User Privacy Gradient Sparsification

November 12, 2023

Preserving Node-level Privacy in Graph Neural Networks
Zihang Xiang, Tianhao Wang, Di Wang
Graph Neural Network Differential Privacy Membership Inference Attack Node Level Privacy

November 10, 2023

Does Differential Privacy Prevent Backdoor Attacks in Practice?
Fereshteh Razmi, Jian Lou, Li Xiong
Differential Privacy Backdoor Attack Practice Mode Differential Privacy SGD

Differential Privacy

Papers

Communication Cost Reduction for Subgraph Counting under Local Differential Privacy via Hash Functions

Classification with Partially Private Features

QMGeo: Differentially Private Federated Learning via Stochastic Quantization with Mixed Truncated Geometric Distribution

Optimal Unbiased Randomizers for Regression with Label Differential Privacy

On the Inadequacy of Similarity-based Privacy Metrics: Reconstruction Attacks against "Truly Anonymous Synthetic Data''

All Rivers Run to the Sea: Private Learning with Asymmetric Flows

Reconciling AI Performance and Data Reconstruction Resilience for Medical Imaging

PAC Privacy Preserving Diffusion Models

Navigating Privacy and Copyright Challenges Across the Data Lifecycle of Generative AI

Using Decentralized Aggregation for Federated Learning with Differential Privacy

Differentially Private SGD Without Clipping Bias: An Error-Feedback Approach

DPSUR: Accelerating Differentially Private Stochastic Gradient Descent Using Selective Update and Release

Differentially Private Non-Convex Optimization under the KL Condition with Optimal Rates

A Theoretical Insight into Attack and Defense of Gradient Leakage in Transformer

Zero redundancy distributed learning with differential privacy

Can we infer the presence of Differential Privacy in Deep Learning models' weights? Towards more secure Deep Learning

Are Normalizing Flows the Key to Unlocking the Exponential Mechanism?

Sparsity-Preserving Differentially Private Training of Large Embedding Models

Preserving Node-level Privacy in Graph Neural Networks

Does Differential Privacy Prevent Backdoor Attacks in Practice?