🚀 We're looking for ML Engineers and Medical Reviewers! Join the OpenPHR Mission →
Back to Marketplace
Dataset

MedNLI

Clinical Text / NLP General PhysioNet Credentialed Restricted Access
N/A GitHub Stars
N/A Open Issues
N/A Docker Support
N/A Last Updated

Technical Summary

MedNLI is a dataset annotated by doctors for Natural Language Inference (NLI) in the clinical domain. It is derived from the clinical notes in MIMIC-III.

Key Capabilities

  • Clinical Reasoning Evaluation: Provides a robust benchmark for evaluating how well a model understands clinical language and logic.
  • Expert Annotated: The dataset was annotated by board-certified physicians, ensuring high-quality, clinically accurate logical relationships between premises and hypotheses.
  • Integration with MIMIC: As an extension of MIMIC-III, it benefits from the vast surrounding ecosystem of open clinical data research.

Model Card Details

Architecture

N/A

Intended Use Cases

Natural Language Inference (NLI) in the clinical domain; determining if a clinical hypothesis is true, false, or undetermined given a clinical premise.