This research introduces P-TA, a novel approach that combines Large Language Models with Proximal Policy Optimization to generate high-quality synthetic tabular data.

• Addresses limitations in existing GAN and LLM-based approaches by incorporating reinforcement learning
• Significantly improves data quality by reducing common-sense errors and better capturing feature relationships
• Demonstrates superior performance across multiple domains including security applications
• Enables more realistic synthetic data generation for testing security systems and fraud detection scenarios

For security professionals, this advancement offers more reliable synthetic datasets for training detection systems, conducting penetration testing, and developing more robust security protocols without compromising sensitive real-world data.

P-TA: Using Proximal Policy Optimization to Enhance Tabular Data Augmentation via Large Language Models