Striking the Perfect Balance: Preserving Privacy While Boosting Utility in Collaborative Medical Prediction Platforms

2025-07-15

AI Summary

Online medical prediction platforms face dual challenges—patient attribute leakage and physician model exposure—leading to degraded predictive performance. Method: This paper proposes the first distributed learning framework that jointly safeguards patient attribute privacy and physician model security. It formally defines and distinguishes two types of privacy attacks (patient-targeted and physician-targeted), and introduces a one-shot distributed learning mechanism. Grounded in statistical learning theory, the framework integrates differential privacy, federated learning, and secure aggregation—requiring no sharing of raw data or model parameters. Contribution/Results: Theoretically, it proves optimality of predictive performance under a given privacy budget. Empirical evaluation on synthetic and real-world healthcare datasets demonstrates robust resistance to both privacy threats, achieves prediction accuracy comparable to centralized training, and significantly enhances platform utility and security.

Technology Category

• Machine Learning: Privacy
• Search and Optimization: Distributed Search
• Natural Language Processing: Safety and Robustness

Application Category

• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Online collaborative medical prediction platforms offer convenience and real-time feedback by leveraging massive electronic health records. However, growing concerns about privacy and low prediction quality can deter patient participation and doctor cooperation. In this paper, we first clarify the privacy attacks, namely attribute attacks targeting patients and model extraction attacks targeting doctors, and specify the corresponding privacy principles. We then propose a privacy-preserving mechanism and integrate it into a novel one-shot distributed learning framework, aiming to simultaneously meet both privacy requirements and prediction performance objectives. Within the framework of statistical learning theory, we theoretically demonstrate that the proposed distributed learning framework can achieve the optimal prediction performance under specific privacy requirements. We further validate the developed privacy-preserving collaborative medical prediction platform through both toy simulations and real-world data experiments.

Problem

• Addressing privacy attacks in medical prediction platforms
• Balancing privacy and utility in collaborative healthcare
• Ensuring optimal prediction performance with privacy constraints

Innovation

• Privacy-preserving mechanism for medical platforms
• One-shot distributed learning framework
• Optimal performance under privacy constraints