Scholarlink

<p>The growing adoption of artificial intelligence in healthcare highlights the need for models that can leverage heterogeneous patient data while preserving strict privacy requirements. This paper proposes a novel multi-modal federated learning framework with differential privacy for decentralized healthcare AI. The model integrates electronic health records and ECG time-series using modality-specific encoders and a shared latent fusion network, enabling comprehensive representation learning without centralizing sensitive data. Differential privacy is incorporated into local updates to provide formal guarantees against information leakage in federated aggregation. Extensive experiments on real-world healthcare datasets show that the proposed method achieves <InlineEquation ID="IEq1"><EquationSource Format="TEX">\(94.12\%\)</EquationSource></InlineEquation> accuracy, <InlineEquation ID="IEq2"><EquationSource Format="TEX">\(93.64\%\)</EquationSource></InlineEquation> precision, <InlineEquation ID="IEq3"><EquationSource Format="TEX">\(93.21\%\)</EquationSource></InlineEquation> recall, <InlineEquation ID="IEq4"><EquationSource Format="TEX">\(93.42\%\)</EquationSource></InlineEquation> F1-score, and <InlineEquation ID="IEq5"><EquationSource Format="TEX">\(95.03\%\)</EquationSource></InlineEquation> AUC, outperforming centralized, single-modality, and non-private baselines. The framework also converges <InlineEquation ID="IEq6"><EquationSource Format="TEX">\(32.4\%\)</EquationSource></InlineEquation> faster than single-modality federated learning, reaching <InlineEquation ID="IEq7"><EquationSource Format="TEX">\(90\%\)</EquationSource></InlineEquation> accuracy in 35 rounds. An ablation study confirms the contribution of multi-modal fusion and class balancing, while client variance analysis shows the lowest performance deviation (<InlineEquation ID="IEq8"><EquationSource Format="TEX">\(\pm 1.2\%\)</EquationSource></InlineEquation>) under heterogeneous distributions. These results indicate that combining federated optimization, differential privacy, and multi-modal learning provides an effective framework for privacy-preserving clinical AI, with potential for deployment in distributed healthcare settings.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-modal federated learning with differential privacy for privacy-preserving healthcare AI

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月9日

查看全文
  • Md. Rokibul Hasan,
  • Md. Istiaq Ahmed,
  • Sudip Saha,
  • Tashnim Khan Ishika,
  • Hashibul Ahsan Shoaib,
  • Md. Jakir Hossen

摘要

The growing adoption of artificial intelligence in healthcare highlights the need for models that can leverage heterogeneous patient data while preserving strict privacy requirements. This paper proposes a novel multi-modal federated learning framework with differential privacy for decentralized healthcare AI. The model integrates electronic health records and ECG time-series using modality-specific encoders and a shared latent fusion network, enabling comprehensive representation learning without centralizing sensitive data. Differential privacy is incorporated into local updates to provide formal guarantees against information leakage in federated aggregation. Extensive experiments on real-world healthcare datasets show that the proposed method achieves \(94.12\%\) accuracy, \(93.64\%\) precision, \(93.21\%\) recall, \(93.42\%\) F1-score, and \(95.03\%\) AUC, outperforming centralized, single-modality, and non-private baselines. The framework also converges \(32.4\%\) faster than single-modality federated learning, reaching \(90\%\) accuracy in 35 rounds. An ablation study confirms the contribution of multi-modal fusion and class balancing, while client variance analysis shows the lowest performance deviation (\(\pm 1.2\%\)) under heterogeneous distributions. These results indicate that combining federated optimization, differential privacy, and multi-modal learning provides an effective framework for privacy-preserving clinical AI, with potential for deployment in distributed healthcare settings.