Accelerating Into the Future: DAF-MIT AI Accelerator Extends for Another Five Years

  • Published
  • By DAF AI Accelerator Public Affairs
  • DAF-MIT AI Accelerator
The Department of the Air Force and the Massachusetts Institute of Technology renewed their cooperative agreement, extending the program known as the DAF-MIT AI Accelerator for another five years. The AI Accelerator advances fundamental research in artificial intelligence that could improve DAF operations while also addressing broader societal needs.
In a ceremony at MIT on May 7, 2024, Chief of Staff of the Air Force Gen. David W. Allvin joined Professor Daniela Rus, director of MIT’s Computer Science and Artificial Intelligence Laboratory and the Andrew and Erna Viterbi Professor of Electrical Engineering and Computer Science, to recognize how the two sides’ strong cooperation and focus on hard problems has brought operationally relevant solutions and expertise to DAF operations by enabling rapid prototyping, scaling, and application of AI algorithms and systems.
“There’s a natural connection between where the DAF is going and the AIA. That’s why we made the conscious decision for the long-term commitment,” Allvin said. “This partnership is something I’m committed to as long as I’m the Chief of Staff of the Air Force.”
The AIA spearheads a novel, highly collaborative, and integrated effort to leverage the expertise and resources of MIT, MIT’s Lincoln Laboratory (a U.S. Department of Defense federally funded research and development center), and other organizations in academia and industry.
“The five-year extension of the AI Accelerator collaboration between the DAF and MIT underscores our joint commitment to advancing the science and engineering of AI, and its applications,” said Rus. “We believe that AI holds enormous potential to improve the quality of every part of our lives and we are especially excited about the potential applications for strengthening national security. In the second phase of the AI Accelerator, we aim for even greater advancements and innovative solutions that will continue to push the boundaries of AI and its ethical applications in addressing societal needs and the defense of the nation. This collaboration is an example of how technology, guided by human insight and ethical considerations, can expand its impact to solve complex problems on a global scale.”
The collaboration supports more than a dozen research projects addressing challenges that are important to both the DAF and society more broadly, such as disaster response and medical readiness.
Under the agreement, MIT will continue to involve interdisciplinary teams of researchers, faculty, and students whose work focuses on topics in artificial intelligence, control theory, formal methods, machine learning, robotics, and perception, among other fields. Teams include leaders in technology from a range of departments, labs, and centers across the Institute. Members of the DAF, to include Airmen and Guardians who are embedded in the projects, will continue to lend expertise to each team and liaise with DAF stakeholders.
“Under current operations, the DAF-MIT AI Accelerator engages with over 150 faculty, researchers, and students, incorporating Airmen and Guardians directly in our research lines of effort,” said Col. Garry “Pink” Floyd, director of the DAF-MIT AI Accelerator. “The accelerator provides and facilitates AI educational opportunities for members across the DAF, and advances science and AI research on hard problems while pursuing solutions that will benefit both the Air Force and humanity. Since the AIA’s inception, nearly 500 papers have been published under our banner, and those papers have been cited over 8000 times. In addition, several of our projects have put operational prototypes into the field for testing and evaluation.”
The AIA includes faculty, staff, and students from all five MIT schools, and is a component of the MIT Stephen A. Schwarzman College of Computing. The college takes an interdisciplinary approach to computing and focuses on the societal implications of computing and AI. MIT Lincoln Laboratory will continue to make available its specialized facilities and resources to support Air Force mission requirements.
Some examples of how fundamental research has helped the DAF from the last five years include:
  • Deploying the Puckboard Intelligent Recommendation Engine prototype at four operational flying squadrons, using novel multi-agent reinforcement learning techniques to build a scheduling assistant that boosts aircrew readiness.
  • Actively implementing Natural Language Processing advancements within the DoD Platform One chat systems at the 618th Air Operations Center, Scott Air Force Base, to enhance command and control with real-time chat data analysis, trend mapping, and information retrieval.
  • Proving that navigation using the Earth’s magnetic field from inside an airplane is possible.
  • Successfully testing Liquid Neural Net-powered autonomous flight, bringing an entire new class of machine learning algorithms closer to operational use.
In addition to disaster relief and medical readiness, other possible research areas may include data management, human-machine teaming advancements, vehicle safety, and cyber resiliency.
“The AI Accelerator provides us with an opportunity to develop technologies that will be vectors for positive change in the world,” Rus says. “Our projects integrate societal implications into research from the outset.”
"The AI Accelerator's research projects aim to develop new algorithms and systems to assist with complex decision-making across a variety of potential applications,” said Floyd, “For example, one of our projects could provide key insights into force structure and other components of long-range strategic planning."