Efficient Differentially Private Fine-Tuning of Diffusion Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Jing, Lowy, Andrew, Koike-Akino, Toshiaki, Parsons, Kieran, Wang, Ye
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910477604356096
author Liu, Jing
Lowy, Andrew
Koike-Akino, Toshiaki
Parsons, Kieran
Wang, Ye
author_facet Liu, Jing
Lowy, Andrew
Koike-Akino, Toshiaki
Parsons, Kieran
Wang, Ye
contents The recent developments of Diffusion Models (DMs) enable generation of astonishingly high-quality synthetic samples. Recent work showed that the synthetic samples generated by the diffusion model, which is pre-trained on public data and fully fine-tuned with differential privacy on private data, can train a downstream classifier, while achieving a good privacy-utility tradeoff. However, fully fine-tuning such large diffusion models with DP-SGD can be very resource-demanding in terms of memory usage and computation. In this work, we investigate Parameter-Efficient Fine-Tuning (PEFT) of diffusion models using Low-Dimensional Adaptation (LoDA) with Differential Privacy. We evaluate the proposed method with the MNIST and CIFAR-10 datasets and demonstrate that such efficient fine-tuning can also generate useful synthetic samples for training downstream classifiers, with guaranteed privacy protection of fine-tuning data. Our source code will be made available on GitHub.
format Preprint
id arxiv_https___arxiv_org_abs_2406_05257
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Efficient Differentially Private Fine-Tuning of Diffusion Models
Liu, Jing
Lowy, Andrew
Koike-Akino, Toshiaki
Parsons, Kieran
Wang, Ye
Machine Learning
Cryptography and Security
The recent developments of Diffusion Models (DMs) enable generation of astonishingly high-quality synthetic samples. Recent work showed that the synthetic samples generated by the diffusion model, which is pre-trained on public data and fully fine-tuned with differential privacy on private data, can train a downstream classifier, while achieving a good privacy-utility tradeoff. However, fully fine-tuning such large diffusion models with DP-SGD can be very resource-demanding in terms of memory usage and computation. In this work, we investigate Parameter-Efficient Fine-Tuning (PEFT) of diffusion models using Low-Dimensional Adaptation (LoDA) with Differential Privacy. We evaluate the proposed method with the MNIST and CIFAR-10 datasets and demonstrate that such efficient fine-tuning can also generate useful synthetic samples for training downstream classifiers, with guaranteed privacy protection of fine-tuning data. Our source code will be made available on GitHub.
title Efficient Differentially Private Fine-Tuning of Diffusion Models
topic Machine Learning
Cryptography and Security
url https://arxiv.org/abs/2406.05257