Selected work

Representative Papers

ATRIA: Adaptive Traceable ECG Reporting with Iterative Agents

Jun 23, 2026

This work proposes a multi-agent framework for ECG report generation that emulates clinicians’ iterative diagnostic reasoning. Unlike existing end-to-end approaches—where errors propagate irreversibly—or agent-based systems lacking revision capabilities, our method explicitly links each diagnostic statement to its supporting evidence, enables dynamic incorporation of new contextual information, and allows clinicians to validate and edit individual findings mid-process. The system introduces, for the first time, a traceable, editable, and bidirectionally iterative reporting mechanism that enhances transparency and aligns with real-world clinical workflows. Built upon a deployed ECG analysis model and a cloud-native architecture, it facilitates efficient human–AI collaboration and demonstrates immediate clinical deployability, as validated through four representative interactive case studies.

0 citationsRead paper

ExECG: An Explainable AI Framework for ECG models

May 18, 2026

Current deep learning models for electrocardiogram (ECG) analysis face significant barriers to clinical deployment due to the absence of a standardized, reproducible framework for explainability. To address this challenge, this work proposes ExECG—the first end-to-end unified explainable AI framework specifically designed for ECG models. ExECG employs a three-layer architecture (Wrapper, Explainer, Visualizer) that supports diverse ECG data formats and integrates multiple explainable AI (XAI) algorithms. By standardizing data ingestion, enforcing a consistent analytical protocol, and providing a uniform visualization interface, the framework ensures interoperability across explanation methods and enhances result reproducibility. Case studies demonstrate that ExECG substantially improves the consistency of model explanations, clinical trustworthiness, and reusability in research settings.

0 citationsRead paper

CoFE: A Framework Generating Counterfactual ECG for Explainable Cardiac AI-Diagnostics

Aug 21, 2025

To address the limited clinical interpretability of AI-based electrocardiogram (ECG) models, this paper proposes CoFE—the first counterfactual explanation framework specifically designed for ECG signals. CoFE integrates temporal modeling with causal counterfactual generation to precisely localize perturbations in clinically meaningful features—such as waveform amplitudes and interval durations—and quantify their causal effects on model predictions (e.g., atrial fibrillation classification or serum potassium level regression). The generated counterfactual ECG signals adhere to domain-specific physiological priors, enabling clinicians to scrutinize and validate model reasoning. Experimental results demonstrate that CoFE substantially enhances model transparency and clinical trustworthiness. It provides a novel, interpretable paradigm for deploying AI-ECG systems in real-world clinical settings.

0 citationsRead paper

ALFRED: Ask a Large-language model For Reliable ECG Diagnosis

Apr 30, 2025

Automatic ECG diagnosis faces challenges of low model reliability, poor interpretability, and heavy dependence on labeled data. Method: This paper proposes a zero-shot retrieval-augmented generation (RAG) framework that—uniquely—integrates an expert-curated, structured cardiology knowledge graph into the RAG pipeline, enabling evidence-driven ECG interpretation in synergy with large language models (LLMs). It combines PTB-XL fine-tuning with medical-knowledge-guided semantic retrieval, eliminating reliance on conventional supervised training. Results: The framework achieves high accuracy on multi-class diagnosis tasks using PTB-XL, demonstrating strong generalization and clinical interpretability: each diagnostic output is accompanied by explicit knowledge provenance and step-by-step reasoning justification. This work establishes a novel paradigm for trustworthy, zero-shot, knowledge-enhanced AI-assisted ECG analysis.

0 citationsRead paper

Recent publications

Latest Papers

ATRIA: Adaptive Traceable ECG Reporting with Iterative Agents

Jun 23, 2026

This work proposes a multi-agent framework for ECG report generation that emulates clinicians’ iterative diagnostic reasoning. Unlike existing end-to-end approaches—where errors propagate irreversibly—or agent-based systems lacking revision capabilities, our method explicitly links each diagnostic statement to its supporting evidence, enables dynamic incorporation of new contextual information, and allows clinicians to validate and edit individual findings mid-process. The system introduces, for the first time, a traceable, editable, and bidirectionally iterative reporting mechanism that enhances transparency and aligns with real-world clinical workflows. Built upon a deployed ECG analysis model and a cloud-native architecture, it facilitates efficient human–AI collaboration and demonstrates immediate clinical deployability, as validated through four representative interactive case studies.

0 citationsRead paper

ExECG: An Explainable AI Framework for ECG models

May 18, 2026

Current deep learning models for electrocardiogram (ECG) analysis face significant barriers to clinical deployment due to the absence of a standardized, reproducible framework for explainability. To address this challenge, this work proposes ExECG—the first end-to-end unified explainable AI framework specifically designed for ECG models. ExECG employs a three-layer architecture (Wrapper, Explainer, Visualizer) that supports diverse ECG data formats and integrates multiple explainable AI (XAI) algorithms. By standardizing data ingestion, enforcing a consistent analytical protocol, and providing a uniform visualization interface, the framework ensures interoperability across explanation methods and enhances result reproducibility. Case studies demonstrate that ExECG substantially improves the consistency of model explanations, clinical trustworthiness, and reusability in research settings.

0 citationsRead paper

CoFE: A Framework Generating Counterfactual ECG for Explainable Cardiac AI-Diagnostics

Aug 21, 2025

To address the limited clinical interpretability of AI-based electrocardiogram (ECG) models, this paper proposes CoFE—the first counterfactual explanation framework specifically designed for ECG signals. CoFE integrates temporal modeling with causal counterfactual generation to precisely localize perturbations in clinically meaningful features—such as waveform amplitudes and interval durations—and quantify their causal effects on model predictions (e.g., atrial fibrillation classification or serum potassium level regression). The generated counterfactual ECG signals adhere to domain-specific physiological priors, enabling clinicians to scrutinize and validate model reasoning. Experimental results demonstrate that CoFE substantially enhances model transparency and clinical trustworthiness. It provides a novel, interpretable paradigm for deploying AI-ECG systems in real-world clinical settings.

0 citationsRead paper

ALFRED: Ask a Large-language model For Reliable ECG Diagnosis

Apr 30, 2025

Automatic ECG diagnosis faces challenges of low model reliability, poor interpretability, and heavy dependence on labeled data. Method: This paper proposes a zero-shot retrieval-augmented generation (RAG) framework that—uniquely—integrates an expert-curated, structured cardiology knowledge graph into the RAG pipeline, enabling evidence-driven ECG interpretation in synergy with large language models (LLMs). It combines PTB-XL fine-tuning with medical-knowledge-guided semantic retrieval, eliminating reliance on conventional supervised training. Results: The framework achieves high accuracy on multi-class diagnosis tasks using PTB-XL, demonstrating strong generalization and clinical interpretability: each diagnostic output is accompanied by explicit knowledge provenance and step-by-step reasoning justification. This work establishes a novel paradigm for trustworthy, zero-shot, knowledge-enhanced AI-assisted ECG analysis.

0 citationsRead paper