Fine-Tuning Large Language Models with User-Level Differential Privacy
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_ | 1866910521508233216 |
|---|---|
| author | Charles, Zachary Ganesh, Arun McKenna, Ryan McMahan, H. Brendan Mitchell, Nicole Pillutla, Krishna Rush, Keith |
| author_facet | Charles, Zachary Ganesh, Arun McKenna, Ryan McMahan, H. Brendan Mitchell, Nicole Pillutla, Krishna Rush, Keith |
| contents | We investigate practical and scalable algorithms for training large language models (LLMs) with user-level differential privacy (DP) in order to provably safeguard all the examples contributed by each user. We study two variants of DP-SGD with: (1) example-level sampling (ELS) and per-example gradient clipping, and (2) user-level sampling (ULS) and per-user gradient clipping. We derive a novel user-level DP accountant that allows us to compute provably tight privacy guarantees for ELS. Using this, we show that while ELS can outperform ULS in specific settings, ULS generally yields better results when each user has a diverse collection of examples. We validate our findings through experiments in synthetic mean estimation and LLM fine-tuning tasks under fixed compute budgets. We find that ULS is significantly better in settings where either (1) strong privacy guarantees are required, or (2) the compute budget is large. Notably, our focus on LLM-compatible training algorithms allows us to scale to models with hundreds of millions of parameters and datasets with hundreds of thousands of users. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_07737 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Fine-Tuning Large Language Models with User-Level Differential Privacy Charles, Zachary Ganesh, Arun McKenna, Ryan McMahan, H. Brendan Mitchell, Nicole Pillutla, Krishna Rush, Keith Machine Learning Computation and Language Cryptography and Security Distributed, Parallel, and Cluster Computing We investigate practical and scalable algorithms for training large language models (LLMs) with user-level differential privacy (DP) in order to provably safeguard all the examples contributed by each user. We study two variants of DP-SGD with: (1) example-level sampling (ELS) and per-example gradient clipping, and (2) user-level sampling (ULS) and per-user gradient clipping. We derive a novel user-level DP accountant that allows us to compute provably tight privacy guarantees for ELS. Using this, we show that while ELS can outperform ULS in specific settings, ULS generally yields better results when each user has a diverse collection of examples. We validate our findings through experiments in synthetic mean estimation and LLM fine-tuning tasks under fixed compute budgets. We find that ULS is significantly better in settings where either (1) strong privacy guarantees are required, or (2) the compute budget is large. Notably, our focus on LLM-compatible training algorithms allows us to scale to models with hundreds of millions of parameters and datasets with hundreds of thousands of users. |
| title | Fine-Tuning Large Language Models with User-Level Differential Privacy |
| topic | Machine Learning Computation and Language Cryptography and Security Distributed, Parallel, and Cluster Computing |
| url | https://arxiv.org/abs/2407.07737 |