Prompt Public Large Language Models to Synthesize Data for Private On-device Applications
This work addresses the challenge of enhancing privacy-preserving on-device applications like mobile keyboards, representing an incremental advance in data synthesis methods.
The paper tackled the problem of improving pre-training data quality for on-device language models trained with differential privacy and federated learning by using large language models to synthesize data that resembles real user distributions, resulting in relative improvements of 19.0% and 22.8% in next word prediction accuracy over a baseline.
Pre-training on public data is an effective method to improve the performance for federated learning (FL) with differential privacy (DP). This paper investigates how large language models (LLMs) trained on public data can improve the quality of pre-training data for the on-device language models trained with DP and FL. We carefully design LLM prompts to filter and transform existing public data, and generate new data to resemble the real user data distribution. The model pre-trained on our synthetic dataset achieves relative improvement of 19.0% and 22.8% in next word prediction accuracy compared to the baseline model pre-trained on a standard public dataset, when evaluated over the real user data in Gboard (Google Keyboard, a production mobile keyboard application). Furthermore, our method achieves evaluation accuracy better than or comparable to the baseline during the DP FL fine-tuning over millions of mobile devices, and our final model outperforms the baseline in production A/B testing. Our experiments demonstrate the strengths of LLMs in synthesizing data close to the private distribution even without accessing the private data, and also suggest future research directions to further reduce the distribution gap.