The USAF-MIT AI Accelerator combines the expertise and resources of MIT and the Air Force and Space Force to conduct fundamental research enabling rapid prototyping, scaling, and application of AI algorithms and systems. The AI Accelerator is designed to make fundamental advances in artificial intelligence that improve Air Force operations while also addressing broader societal needs. AI Accelerator research teams consist of interdisciplinary teams, including Airmen, that collaborate across different fields of AI to create new algorithms and tools. The Air Force plans to invest approximately $15 million per year as it builds upon its five-decade relationship with MIT (MIT News).

The AI Accelerator involves researchers, professors, and students from across MIT Campus and MIT Lincoln Laboratory. More than 10 Airmen are embedded in the Accelerator’s research projects at MIT. AI Accelerator stakeholders include many Air Force units, the Department of the Navy, the Joint Artificial Intelligence Center (JAIC), Air Force Chief Data Office, AFWERX, KesselRun, Air Force Research Laboratory (AFRL), and Air Force Cyberworx.

For Academia & Industry: Researchers can participate in the challenge problems developed by the AI Accelerator research projects.
For Airmen and Space Professionals: Please reach out to us using the Contact Us page if you or your team think that you have challenges that can be solved by AI. We also offer internships for USAF/USSF ROTC and Air Force Academy Cadets.

We accelerate fundamental AI research through rapid prototyping, scaling, and application of AI algorithms and systems to improve Air Force operations and address underserved societal needs.

A Challenge is a means to engage academia, industry, and the public towards solving a common problem. In the context of Machine Learning, challenges have commonly involved open competitions organized around a publicly released data set. The team or individual developing the best Machine Learning model to learn from or analyze that data set wins! The world-wide collaboration resulting from such challenges has historically driven fundamental advances in many fields. In some communities, these challenges are also called “Hackathons.” Challenges such as the Datathon (DoD challenge) and the MagNav (public challenge) at JuliaCon play an important role in advancing diverse fields of AI research. A major goal of each AI Accelerator project is to host challenges for the public.

AI Accelerator publications can be found on our Google Scholar page.