Selected work

Representative Papers

Communication-Efficient Federated Risk Difference Estimation for Time-to-Event Clinical Outcomes

Jan 21, 2026

This study addresses the challenge of absolute survival risk estimation in multi-center medical research, where privacy constraints and reliance on a central server hinder clinically interpretable inference. Existing federated learning approaches typically yield only relative effect measures and lack interpretability for absolute risk. To overcome these limitations, we propose FedRD, a novel server-agnostic and communication-efficient federated framework for estimating risk differences. FedRD requires only one round of communication under stratified settings or three rounds in non-stratified scenarios, enabling confidence interval construction and hypothesis testing directly from distributed survival data. Theoretically, the non-stratified variant is shown to be asymptotically equivalent to centralized analysis. Extensive experiments on both simulated and real-world multinational datasets demonstrate that FedRD substantially outperforms local analyses and existing federated baselines, delivering privacy-preserving, interpretable, and statistically inferable absolute risk estimates.

0 citationsRead paper

Opening the black box of deep learning: Validating the statistical association between explainable artificial intelligence (XAI) and clinical domain knowledge in fundus image-based glaucoma diagnosis

Apr 06, 2025

Deep learning models in medical image diagnosis suffer from limited interpretability, undermining clinical trust. To address this, we conduct the first systematic cross-model, cross-method, and multi-dataset explainable AI (XAI) validation for glaucoma fundus image classification. We evaluate five class activation mapping (CAM) techniques—including Grad-CAM—across four architectures (VGG-11, ResNet-18, DeiT-Tiny, Swin-Tiny) on five publicly available datasets. Attention heatmaps are quantitatively aligned with clinically relevant anatomical structures (optic cup, optic disc, retinal vessels). Paired t-tests and Pearson/Spearman correlation analyses demonstrate that all models significantly attend to these regions (p < 0.001), and the anatomical coverage ratio of attention maps strongly correlates with diagnostic accuracy (r > 0.8, p < 0.001). This work provides the first statistical evidence—spanning multiple models, explanation methods, and datasets—supporting the clinical credibility of XAI in ophthalmic diagnosis.

0 citationsRead paper

Recent publications

Latest Papers

Communication-Efficient Federated Risk Difference Estimation for Time-to-Event Clinical Outcomes

Jan 21, 2026

This study addresses the challenge of absolute survival risk estimation in multi-center medical research, where privacy constraints and reliance on a central server hinder clinically interpretable inference. Existing federated learning approaches typically yield only relative effect measures and lack interpretability for absolute risk. To overcome these limitations, we propose FedRD, a novel server-agnostic and communication-efficient federated framework for estimating risk differences. FedRD requires only one round of communication under stratified settings or three rounds in non-stratified scenarios, enabling confidence interval construction and hypothesis testing directly from distributed survival data. Theoretically, the non-stratified variant is shown to be asymptotically equivalent to centralized analysis. Extensive experiments on both simulated and real-world multinational datasets demonstrate that FedRD substantially outperforms local analyses and existing federated baselines, delivering privacy-preserving, interpretable, and statistically inferable absolute risk estimates.

0 citationsRead paper

Opening the black box of deep learning: Validating the statistical association between explainable artificial intelligence (XAI) and clinical domain knowledge in fundus image-based glaucoma diagnosis

Apr 06, 2025

Deep learning models in medical image diagnosis suffer from limited interpretability, undermining clinical trust. To address this, we conduct the first systematic cross-model, cross-method, and multi-dataset explainable AI (XAI) validation for glaucoma fundus image classification. We evaluate five class activation mapping (CAM) techniques—including Grad-CAM—across four architectures (VGG-11, ResNet-18, DeiT-Tiny, Swin-Tiny) on five publicly available datasets. Attention heatmaps are quantitatively aligned with clinically relevant anatomical structures (optic cup, optic disc, retinal vessels). Paired t-tests and Pearson/Spearman correlation analyses demonstrate that all models significantly attend to these regions (p < 0.001), and the anatomical coverage ratio of attention maps strongly correlates with diagnostic accuracy (r > 0.8, p < 0.001). This work provides the first statistical evidence—spanning multiple models, explanation methods, and datasets—supporting the clinical credibility of XAI in ophthalmic diagnosis.

0 citationsRead paper