Published September 24, 2026 | Version v3
Dataset Open

Synthetic AI-READI Dataset for T2DM Research

Authors/Creators

  • 1. ROR icon Vanderbilt University
  • 2. EDMO icon Princeton University
  • 3. ROR icon Vanderbilt University Medical Center
  • 4. The University of Texas Health Science Center at Houston
  • 5. ROR icon University of Washington

Description

This record archives a synthetic multimodal dataset for type 2 diabetes, derived from the AI-READI dataset (v3.0.0; https://doi.org/10.60775/fairhub.3). The synthetic data comprise three independently generated modules — retinal optical coherence tomography (OCT) images, retinal fundus photographs, and tabular clinical records — formatted to match the structure of the original AI-READI data. Because the data are entirely synthetic, they contain no records corresponding to real individuals and are intended to support method development, benchmarking, education, and reproducible research without the access constraints of the source dataset.

This Zenodo record is the citable archival version of the dataset. The data are additionally available through Amazon Web Services Marketplace: https://aws.amazon.com/marketplace/pp/prodview-mxp3i7s3ugkxq

For details of the dataset's construction and validation, see the associated publication [citation / DOI to be added on acceptance].

 

Files

license (1).pdf

Files (52.7 kB)

Name Size
md5:97133111a2d72e4780bf674722693bc9
52.7 kB Preview Download

Additional details

Funding

United States National Library of Medicine
Vanderbilt Biomedical Informatics Training Program T15LM007450
Office of the Director
Bridge2AI: Salutogenesis Data Generation Project OT2OD032644
National Human Genome Research Institute
Ethics Core (FABRIC) U54HG012510
United States National Library of Medicine
Generative AI for synthetic data: A framework to expand health data reach for research and ensure algorithmic fairness K99LM014428

Software

Repository URL
https://github.com/nicholas-j-jackson/ai_readi_synth
Programming language
Python