DR-Encoder: Encode Low-rank Gradients with Random Prior for Large Language Models Differentially Privately
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914160655204352 |
|---|---|
| author | Wu, Huiwen Zhang, Deyi Li, Xiaohan Xu, Xiaogang Wu, Jiafei Liu, Zhe |
| author_facet | Wu, Huiwen Zhang, Deyi Li, Xiaohan Xu, Xiaogang Wu, Jiafei Liu, Zhe |
| contents | The emergence of the Large Language Model (LLM) has shown their superiority in a wide range of disciplines, including language understanding and translation, relational logic reasoning, and even partial differential equations solving. The transformer is the pervasive backbone architecture for the foundation model construction. It is vital to research how to adjust the Transformer architecture to achieve an end-to-end privacy guarantee in LLM fine-tuning. In this paper, we investigate three potential information leakage during a federated fine-tuning procedure for LLM (FedLLM). Based on the potential information leakage, we provide an end-to-end privacy guarantee solution for FedLLM by inserting two-stage randomness. The first stage is to train a gradient auto-encoder with a Gaussian random prior based on the statistical information of the gradients generated by local clients. The second stage is to fine-tune the overall LLM with a differential privacy guarantee by adopting appropriate Gaussian noises. We show the efficiency and accuracy gains of our proposed method with several foundation models and two popular evaluation benchmarks. Furthermore, we present a comprehensive privacy analysis with Gaussian Differential Privacy (GDP) and Renyi Differential Privacy (RDP). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_17053 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | DR-Encoder: Encode Low-rank Gradients with Random Prior for Large Language Models Differentially Privately Wu, Huiwen Zhang, Deyi Li, Xiaohan Xu, Xiaogang Wu, Jiafei Liu, Zhe Machine Learning Artificial Intelligence Cryptography and Security The emergence of the Large Language Model (LLM) has shown their superiority in a wide range of disciplines, including language understanding and translation, relational logic reasoning, and even partial differential equations solving. The transformer is the pervasive backbone architecture for the foundation model construction. It is vital to research how to adjust the Transformer architecture to achieve an end-to-end privacy guarantee in LLM fine-tuning. In this paper, we investigate three potential information leakage during a federated fine-tuning procedure for LLM (FedLLM). Based on the potential information leakage, we provide an end-to-end privacy guarantee solution for FedLLM by inserting two-stage randomness. The first stage is to train a gradient auto-encoder with a Gaussian random prior based on the statistical information of the gradients generated by local clients. The second stage is to fine-tune the overall LLM with a differential privacy guarantee by adopting appropriate Gaussian noises. We show the efficiency and accuracy gains of our proposed method with several foundation models and two popular evaluation benchmarks. Furthermore, we present a comprehensive privacy analysis with Gaussian Differential Privacy (GDP) and Renyi Differential Privacy (RDP). |
| title | DR-Encoder: Encode Low-rank Gradients with Random Prior for Large Language Models Differentially Privately |
| topic | Machine Learning Artificial Intelligence Cryptography and Security |
| url | https://arxiv.org/abs/2412.17053 |