DISCO-TAB: A Hierarchical Reinforcement Learning Framework for Privacy-Preserving Synthesis of Complex Clinical Data

2026-04-01

AI Summary

This work addresses the scarcity of high-quality, privacy-preserving biomedical data that hinders clinical decision support systems, as existing generative models struggle to capture the complex nonlinear dependencies and severe class imbalance inherent in electronic health records. To overcome these challenges, the authors propose DISCO-TAB, a framework that integrates a fine-tuned large language model with a multi-objective discriminator. It employs hierarchical reinforcement learning across four granularities—token, sentence, feature, and row—to deliver fine-grained feedback, while incorporating automatic constraint discovery and inverse-frequency reward shaping to prevent minority-class collapse and preserve clinical logic. Evaluated on high-dimensional, small-sample datasets such as those for heart failure and Parkinson’s disease, the method substantially improves downstream classifier performance (up to +38.2%) while maintaining high statistical fidelity (JSD < 0.01) and strong privacy guarantees.

Abstract

The development of robust clinical decision support systems is frequently impeded by the scarcity of high-fidelity, privacy-preserving biomedical data. While Generative Large Language Models (LLMs) offer a promising avenue for synthetic data generation, they often struggle to capture the complex, non-linear dependencies and severe class imbalances inherent in Electronic Health Records (EHR), leading to statistically plausible but clinically invalid records. To bridge this gap, we introduce DISCO-TAB (DIScriminator-guided COntrol for TABular synthesis), a novel framework that orchestrates a fine-tuned LLM with a multi-objective discriminator system optimized via Reinforcement Learning. Unlike prior methods relying on scalar feedback, DISCO-TAB evaluates synthesis at four granularities, token, sentence, feature, and row, while integrating Automated Constraint Discovery and Inverse-Frequency Reward Shaping to autonomously preserve latent medical logic and resolve minority-class collapse. We rigorously validate our framework across diverse benchmarks, including high-dimensional, small-sample medical datasets (e.g., Heart Failure, Parkinson's). Our results demonstrate that hierarchical feedback yields state-of-the-art performance, achieving up to 38.2% improvement in downstream clinical classifier utility compared to GAN and Diffusion baselines, while ensuring exceptional statistical fidelity (JSD < 0.01) and robust resistance to membership inference attacks. This work establishes a new standard for generating trustworthy, utility-preserving synthetic tabular data for sensitive healthcare applications.

Problem

• privacy-preserving
• synthetic clinical data
• Electronic Health Records
• class imbalance
• clinical validity

Innovation

• Hierarchical Reinforcement Learning
• Privacy-Preserving Synthetic Data
• Tabular Data Generation
• Automated Constraint Discovery
• Inverse-Frequency Reward Shaping