paper-with-me

홈 › Papers

On the Convergence of Differentially-Private Fine-tuning: To Linearly Probe or to Fully Fine-tune?

2024-02-29 · Shuqi Ke, Charlie Hou, Giulia Fanti, Sewoong Oh

Differentially private (DP) machine learning pipelines typically involve a two-phase process: non-private pre-training on a public dataset, followed by fine-tuning on private data using DP optimization techniques. In the DP setting, it has been observed that full fine-tuning may not always yield the best test accuracy, even for in-distribution data. This paper (1) analyzes the training dynamics of DP linear probing (LP) and full fine-tuning (FT), and (2) explores the phenomenon of sequential fine-tuning, starting with linear probing and transitioning to full fine-tuning (LP-FT), and its impact on test loss. We provide theoretical insights into the convergence of DP fine-tuning within an overparameterized neural network and establish a utility curve that determines the allocation of privacy budget between linear probing and full fine-tuning. The theoretical results are supported by empirical evaluations on various benchmarks and models. The findings reveal the complex nature of DP fine-tuning methods. These results contribute to a deeper understanding of DP machine learning and highlight the importance of considering the allocation of privacy budget in the fine-tuning process.

📄 PDF Abstract BibTeX arXiv:2402.18905

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

DP-FedAdamW: An Efficient Optimizer for Differentially Private Federated Large Models

2026-02-23 · Jin Liu, Yinbin Miao, Ning Xi, Junkang Liu arxiv

Balancing convergence efficiency and robustness under Differential Privacy (DP) is a central challenge in Federated Learning (FL). While AdamW accelerates training and fine-tuning in large-scale models, we find that dire…

Federated Learning

Differentially Private ADMM for Distributed Medical Machine Learning

2019-01-07 · Jiahao Ding, Xiaoqi Qin, Wenjun Xu, Yanmin Gong 외

Due to massive amounts of data distributed across multiple locations, distributed machine learning has attracted a lot of research interests. Alternating Direction Method of Multipliers (ADMM) is a powerful method of des…

BIG-bench Machine Learning

Learning to be adversarially robust and differentially private

2022-01-06 · Jamie Hayes, Borja Balle, M. Pawan Kumar

We study the difficulties in learning that arise from robust and differentially private optimization. We first study convergence of gradient descent based adversarial training with differential privacy, taking a simple b…

Binary Classification

Differentially Private Fine-tuning of Language Models

2021-10-13 · ICLR 2022 4 · Da Yu, Saurabh Naik, Arturs Backurs, Sivakanth Gopi 외

We give simpler, sparser, and faster algorithms for differentially private fine-tuning of large-scale pre-trained language models, which achieve the state-of-the-art privacy versus utility tradeoffs on many standard NLP …

Text Generation

Optimal Locally Private Nonparametric Classification with Public Data

2023-11-19 · Yuheng Ma, Hanfang Yang

In this work, we investigate the problem of public data assisted non-interactive Local Differentially Private (LDP) learning with a focus on non-parametric classification. Under the posterior drift assumption, we for the…

Classification