NAII AI Tech Sprints
AI Tech Sprints provide a novel approach to innovation to meet Veteran needs. AI Tech Sprints are three-month competitive engagements that foster collaboration between industry, academia, and the Department of Veterans Affairs (VA). Teams compete to create AI-enabled tools that leverage federal data to address specific Veteran health care issues.
The National Artificial Intelligence Institute (NAII) leads AI Tech Sprints by making federal data available and incentivizing collaboration and innovation through the Government Innovation Framework, Challenge.Gov. Industry, academia, and nonprofit organizations are encouraged to work together, with input from VA researchers and clinicians. Sprint participants have access to several federal data sets, including synthetic VA data.
AI Tech Sprints result in innovations that address real-world health care challenges faced by Veterans.
- Gil Alterovitz, PhD, NAII Director
Sprint partners iteratively design an intervention using federal and private data. Teams receive access to:
- large unique federal data sources (e.g., clinical trials, patents, experimental therapeutics, patients)
- Veteran and expert perspectives
- technical AI and machine learning support
- feedback on demos and from users
- longer-term partnership and/or funding opportunities
Top winners of the AI Tech Sprints receive prize money. Recognized prototypes may be invited to conduct a pilot of the product within VA or be introduced to another agency for next steps in product development.
Learn more about AI Tech Sprints!
Featured in Gov CIO: Listen: Breaking Down AI Tech Sprints at Veterans Affairs
AI Tech Sprint 2022: ASPIRE Demonstrator
Teams will work with cross-agency educators and AI experts to create or refine a computer-adaptive skill assessment platform which will be used across the federal government to train and assess the workforce of the future.
Teams will produce a prototype learning platform for the All Service Personnel Readiness Engine (ASPIRE) project. The general goal of the platform is to administer a computer-adaptive assessment of the test-taker’s skills and understanding of AI. The system will provide an assessment summary and make personalized recommendations of courses or other materials to advance the test-taker’s skills. The software should provide a relatively stable and secure experience, with auditability and interpretability of results. A more detailed description of these goals will be provided during the Tech Sprint sessions. Results will be presented to a panel of experts who will assign a numeric score. Aspects of the software will also be beta-tested and evaluated based on quantitative performance metrics.
UPDATED KEY DATES:
- Applications due: September 16, 2022
- Teams selected by: September 20, 2022
- Tech Sprint kickoff meeting: September 23, 2022
- Finalists selected: December 14, 2022 (estimated)
- Tech Sprint end: December 19, 2022 (estimated)
- Demo Day: January 17, 2023 (estimated)
Need more information?
AI Tech Sprints 2020-21 Winners
Congratulations to our winners!
The 2020-21 Tech Sprint theme was Interventions for Veterans Not Currently Served by VA. More than 61 teams signed up for the 12-week sprint, an increase of 400% from the previous Sprint. The winners cumulatively received $100,000 in prize money awarded in coordination with the General Services Administration’s Challenge.Gov platform.
The top three winners:
Behavidence won first place and a $50,000 prize for developing a smartphone application that monitors Veteran activity, categorizes users by similar behavior, and flags for follow-up of those at increased risk for suicide.
SoKat Consulting, LLC, scored second and received a $25,000 prize for creating a chatbot that can integrate with VA’s Blue Button medical records access. The chatbot can help Veterans get answers to questions and better understand their health care between visits.
General Dynamics IT won third place and $10,000 for an algorithm that can classify skin lesions and help medical staff determine if the quality of an image is good enough to make a skin cancer diagnosis.