Pilot/Opportunity Support for Generative Pretrained Transformers (GPT) to Accelerate Diabetes Research

May 2024 Council

Lead Division/Office

DEM

Point(s) of Contact

Xujing Wang, Ph.D.

Executive Summary

One major challenge in Type 2 Diabetes (T2D) is the enormous heterogeneity associated with the disease. To address this challenge, the research field has generated a large amount of data, and has accumulated a vast amount of prior knowledge about the disease. The data and prior knowledge contain critical, but mostly hidden, information relevant to solving this challenge, however, cross cutting gaps exist in integrating them and extracting predictive signals. One of the major problems is the lack of data science expertise in the T2D research field, and the lack of T2D-specific data science expert systems, models, and tools. This initiative proposes to address this problem by establishing a pilot funding program that leverages the emerging opportunities from recent data science advances. It will recruit multidisciplinary teams that include both T2D and data science experts to (1) develop AI foundation models for T2D; (2) validate the models with top research questions in T2D heterogeneity; (3) disseminate the models and engage the community for further development, validation and application; and (4) develop use cases that demonstrate the models’ potential in accelerating the tempo of research. The expected outcomes will include new and diverse AI experts that have joined T2D research workforce, new AI models that other T2D researchers can use, and informative use cases of the new models.