Private Seeds, Public LLMs: Realistic and Privacy-Preserving Synthetic Data Generation
It addresses privacy concerns for users handling sensitive text data, but appears incremental as it builds on existing private synthetic data generation techniques.
The paper tackled the problem of generating synthetic replicas of private text by balancing privacy and utility, proposing RPSG which uses differential privacy and private seeds to achieve high fidelity and strong privacy protection in experiments against state-of-the-art methods.
Large language models (LLMs) have emerged as a powerful tool for synthetic data generation. A particularly important use case is producing synthetic replicas of private text, which requires carefully balancing privacy and utility. We propose Realistic and Privacy-Preserving Synthetic Data Generation (RPSG), which leverages privacy-preserving mechanisms, including formal differential privacy (DP); and private seeds, in particular text containing personal information, to generate realistic synthetic data. Comprehensive experiments against state-of-the-art private synthetic data generation methods demonstrate that RPSG achieves high fidelity to private data while providing strong privacy protection.