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

Abstract

Differential Privacy has become a widely popular method for data protection in machine learning, especially since it allows formulating strict mathematical privacy guarantees. This survey provides an overview of the state-of-the-art of differentially private centralized deep learning, thorough analyses of recent advances and open problems, as well as a discussion of potential future developments in the field. Based on a systematic literature review, the following topics are addressed: auditing and evaluation methods for private models, improvements of privacy-utility trade-offs, protection against a broad range of threats and attacks, differentially private generative models, and emerging application domains.

arXiv comment35 pages, 2 figures

References in corpus (28)

• Explaining and Harnessing Adversarial Examples
• Deep Learning with Differential Privacy
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
• Renyi Differential Privacy
• Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoning
• Differentially Private Generative Adversarial Network
• Differentially Private Model Publishing for Deep Learning
• Semi-supervised Knowledge Transfer for Deep Learning from Private Training Data
• More Than Privacy: Applying Differential Privacy in Key Areas of Artificial Intelligence
• When Machine Unlearning Jeopardizes Privacy
• Privacy Preserving Face Recognition Utilizing Differential Privacy
• Differentially Private Synthetic Medical Data Generation using Convolutional GANs
• Privacy in Deep Learning: A Survey
• Differential Private Knowledge Transfer for Privacy-Preserving Cross-Domain Recommendation
• A Critical Review on the Use (and Misuse) of Differential Privacy in Machine Learning
• PPGAN: Privacy-preserving Generative Adversarial Network
• Why gradient clipping accelerates training: A theoretical justification for adaptivity
• Large Language Models Can Be Strong Differentially Private Learners
• Robust Anomaly Detection and Backdoor Attack Detection Via Differential Privacy
• Differentially Private Learning Needs Better Features (or Much More Data)
• Survey on Causal-based Machine Learning Fairness Notions
• Guidelines for Implementing and Auditing Differentially Private Systems
• Debugging Differential Privacy: A Case Study for Privacy Auditing
• Position: Considerations for Differentially Private Learning with Large-Scale Public Pretraining
• An Empirical Analysis of Fairness Notions under Differential Privacy
• Differential Privacy Meets Neural Network Pruning
• Synthetic Text Generation with Differential Privacy: A Simple and Practical Recipe

Cited by in corpus (4)

• Edge-Cloud Collaborative Computing on Distributed Intelligence and Model Optimization: A Survey
• Differential privacy for medical deep learning: methods, tradeoffs, and deployment implications
• PPFL-RDSN: Privacy-Preserving Federated Learning-based Residual Dense Spatial Networks for Encrypted Lossy Image Reconstruction
• DP-CDA: An Algorithm for Enhanced Privacy Preservation in Dataset Synthesis Through Randomized Mixing