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

arXiv cs.LG / 2026/4/3

💬 オピニオンSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

要点

DISCO-TAB is a new hierarchical reinforcement learning framework that fine-tunes an LLM and uses a multi-objective discriminator system to generate privacy-preserving synthetic clinical tabular data from EHRs.
The method evaluates generation at multiple feedback granularities (token, sentence, feature, and row) to better capture complex non-linear dependencies and address severe class imbalance that can otherwise produce clinically invalid but statistically plausible records.
It incorporates Automated Constraint Discovery and Inverse-Frequency Reward Shaping to preserve latent medical logic and mitigate minority-class collapse during synthesis.
Experiments on small-sample, high-dimensional medical datasets (e.g., Heart Failure and Parkinson’s) show up to a 38.2% improvement in downstream clinical classifier utility over GAN and diffusion baselines, with strong statistical fidelity (JSD < 0.01) and resistance to membership inference attacks.
The authors position DISCO-TAB as a step toward a new standard for trustworthy synthetic healthcare data generation that maintains both utility and 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.