DP-SGD for Fine-Tuning Foundation Models: A Privacy-Utility Trade-off Study

Yu-Xiang Wang, Borja Balle, Shiva Prasad Kasiviswanathan

Abstract

Investigates applying differentially private stochastic gradient descent to fine-tune large foundation models, characterizing the privacy-utility trade-off.

Categories

Tags

Framework Mappings

Cite This Resource

@article{llmsec202400029,
  title = {DP-SGD for Fine-Tuning Foundation Models: A Privacy-Utility Trade-off Study},
  author = {Yu-Xiang Wang and Borja Balle and Shiva Prasad Kasiviswanathan},
  year = {2024},
  journal = {ICLR 2024},
  url = {https://arxiv.org/abs/2402.06677},
}