Differential Privacy Regularization: Protecting Training Data Through Loss Function Regularization

Francisco Aguilera-Martínez; Fernando Berzal

Differential Privacy Regularization: Protecting Training Data Through Loss Function Regularization

Machine Learning 2024-09-26 v1 Artificial Intelligence Cryptography and Security Neural and Evolutionary Computing

Authors: Francisco Aguilera-Martínez , Fernando Berzal

Abstract

Training machine learning models based on neural networks requires large datasets, which may contain sensitive information. The models, however, should not expose private information from these datasets. Differentially private SGD [DP-SGD] requires the modification of the standard stochastic gradient descent [SGD] algorithm for training new models. In this short paper, a novel regularization strategy is proposed to achieve the same goal in a more efficient manner.

Keywords

differential privacy distributed learning regularization

Cite

@article{arxiv.2409.17144,
  title  = {Differential Privacy Regularization: Protecting Training Data Through Loss Function Regularization},
  author = {Francisco Aguilera-Martínez and Fernando Berzal},
  journal= {arXiv preprint arXiv:2409.17144},
  year   = {2024}
}

Related papers

View all related →