This Challenge seeks meaningful, actionable, innovative, and practical strategies to foster early community engagement and build new academic-community partnerships for future research aligned with the NIDDK mission.

Start Registration

Participants will build an end-to-end AI system that responds to complex biomedical questions by synthesizing evidence from PubMed® and PubMed Central® (PMC). Solutions must provide concise, objective, evidence-grounded responses with valid citations and relevant passages.

Start Registration

Given a challenge-specific corpus of documents comprised of documents selected from PubMed® and PubMed Central ® (PMC), participants are tasked with retrieving the specific, relevant document that satisfy a question/request expressed in natural language.

Start Registration

dbGaP contains vast genomic and phenotypic metadata written inconsistently across studies. This task asks solvers to build AI systems that map raw, unstructured variables to standardized ontology concepts, enabling precise, interoperable study discovery and cohort evaluation.

Start Registration

This task requires participants to create a privacy‑preserving conversational AI that uses public dbGaP documentation, metadata, and aggregate statistics to answer complex cohort‑feasibility questions, helping researchers verify study fit before submitting potentially Data Access Requests.

Start Registration

In September 2023, NIDDK-CR announced the Data Centric Challenge aimed at enhancing NIDDK datasets for future AI applications. Challenge participants were tasked with generating an AI-ready dataset that could be used for future data challenges and producing methods to enhance the AI-readiness of NIDDK data. Participation in the Challenge was tiered (i.e., beginner-level and intermediate/advanced-level) and utilized data from two longitudinal studies focused on type 1 diabetes (TEDDY and TrialNet).