ECG-Protocol-Guided-Grounding-CoT: Protocol-Guided Multimodal LLM for ECG Interpretation

ECG-R1 is a dataset for training and evaluating multimodal large language models on electrocardiography interpretation. It was created by researchers including Jiarui Jin, Haoyu Wang, and others from PKUDigitalHealth. The dataset was last updated on February 9, 2026.

Use Cases

• Training protocol-guided MLLMs for ECG analysis based on the described methodology.
• Benchmarking the reliability of multimodal AI in clinical diagnostics based on the description.
• Developing modality-agnostic models for medical signal interpretation based on the project's goal.

Strengths

• Designed specifically for reliable ECG interpretation, addressing a noted gap in existing MLLMs.
• Created by a team of researchers from PKUDigitalHealth and other institutions.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

PKUDigitalHealth

Freshness

Last updated 2026-02-09 07:22:22; freshness should be verified.