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

October 18, 2023

Federated Heterogeneous Graph Neural Network for Privacy-preserving Recommendation
Bo Yan, Yang Cao, Haoyu Wang, Wenchuan Yang, Junping Du, Chuan Shi
Differential Privacy Recommender System Effective Recommendation Recommendation Model Heterogeneous Graph Neural Network Heterogeneous Information Network Heterogeneous Recommendation

October 17, 2023

Disentangling the Linguistic Competence of Privacy-Preserving BERT
Stefan Arnold, Nils Kemmerzell, Annika Schreiner
Language Model Differential Privacy Internal Representation Representational Similarity Text Privatization Linguistic Competence

October 15, 2023

Private Synthetic Data Meets Ensemble Learning
Haoyuan Sun, Navid Azizan, Akash Srivastava, Hao Wang
Synthetic Data Differential Privacy Ensemble Learning Synthetic Dataset Private Generative Adversarial Network

October 13, 2023

User Inference Attacks on Large Language Models
Nikhil Kandpal, Krishna Pillutla, Alina Oprea, Peter Kairouz, Christopher A. Choquette-Choo, Zheng Xu
Large Language Model Differential Privacy Fine Tuned Large Language Model Inference Attack User Representation

October 10, 2023

October 5, 2023

October 4, 2023

Practical, Private Assurance of the Value of Collaboration via Fully Homomorphic Encryption
Hassan Jameel Asghar, Zhigang Lu, Zhongrui Zhao, Dali Kaafar
Machine Learning Differential Privacy Homomorphic Encryption Net Present Value Practical Application Communication Protocol

October 3, 2023

FLEDGE: Ledger-based Federated Learning Resilient to Inference and Backdoor Attacks
Jorge Castillo, Phillip Rieger, Hossein Fereidooni, Qian Chen, Ahmad Sadeghi
Differential Privacy Federated Learning Backdoor Attack Scientific Inference Reward Function Secure Multi Party Computation

September 29, 2023

Federated Learning with Differential Privacy for End-to-End Speech Recognition
Martin Pelikan, Sheikh Shams Azam, Vitaly Feldman, Jan "Honza" Silovsky, Kunal Talwar, Tatiana Likhomanenko
Automatic Speech Recognition Differential Privacy End to End Recognition Benchmark Large Transformer Model

September 28, 2023

September 21, 2023

September 19, 2023

September 17, 2023

Differential Privacy

Papers

Federated Heterogeneous Graph Neural Network for Privacy-preserving Recommendation

Disentangling the Linguistic Competence of Privacy-Preserving BERT

Private Synthetic Data Meets Ensemble Learning

User Inference Attacks on Large Language Models

Federated Quantum Machine Learning with Differential Privacy

Differentially Private Multi-Site Treatment Effect Estimation

PrIeD-KIE: Towards Privacy Preserved Document Key Information Extraction

FLAIM: AIM-based Synthetic Data Generation in the Federated Setting

Practical, Private Assurance of the Value of Collaboration via Fully Homomorphic Encryption

FLEDGE: Ledger-based Federated Learning Resilient to Inference and Backdoor Attacks

Federated Learning with Differential Privacy for End-to-End Speech Recognition

Recent Advances of Differential Privacy in Centralized Deep Learning: A Systematic Survey

Differentially Private Secure Multiplication: Hiding Information in the Rubble of Noise

User-Level Differential Privacy With Few Examples Per User

S-BDT: Distributed Differentially Private Boosted Decision Trees

Privacy-Preserving In-Context Learning with Differentially Private Few-Shot Generation

DPpack: An R Package for Differentially Private Statistical Analysis and Machine Learning

Striking a Balance: An Optimal Mechanism Design for Heterogenous Differentially Private Data Acquisition for Logistic Regression

Private Matrix Factorization with Public Item Features

Towards Differential Privacy in Sequential Recommendation: A Noisy Graph Neural Network Approach