AFRL teams with academia to win Phase I of AFRL Grand Challenge Published May 8, 2023 By Gail L. Forbes Air Force Research Laboratory WRIGHT-PATTERSON AIR FORCE BASE, Ohio -- The Air Force Research Laboratory, or AFRL, has selected a joint research team from Carnegie Mellon University and the University of North Carolina at Chapel Hill as the Phase I winner of AFRL’s Active Artificial Intelligence, or AI, Planners for Chemistry/Materials Optimization and Discovery Grand Challenge. The winning team proposed a solution to accelerate chemistry and materials research to advance warfighting capabilities for the U.S. Department of the Air Force and the Department of Defense. Program managers within AFRL’s Materials and Manufacturing Directorate, who identified this year’s Grand Challenge topic, wrote the initial call for proposals and ultimately selected the winning team, will oversee the team’s research efforts and provide technical direction over the course of the nine-month challenge effort. Participants in this year’s AFRL headquarters-sponsored Grand Challenge pitched ideas for the development of a machine learning framework, a system that uses artificially intelligent computer systems to support solutions for optimizing and discovering synthetic compounds — manmade substances produced via chemical reactions that have a range of applications for consumer goods, from foods to fuel, as well as myriad defense sector needs. The Air Force Research Laboratory, or AFRL, has selected a joint research team from Carnegie-Mellon University and the University of North Carolina as the Phase I winner of the AFRL headquarters-sponsored Grand Challenge, an opportunity for small businesses, startups and academic teams to propose potential solutions to meet wide-ranging U.S. Department of the Air Force warfighter needs. The winning team pitched a solution for a machine learning-artificial intelligence system that will support the optimization and discovery of synthetic compounds, manmade substances that are applicable to a wide range of defense sector needs. Machine learning, a subfield of artificial intelligence that gives computers the ability to learn from experience and operate without explicit programming or instructions, has significant future implications for a wide range of academic and industrial fields, including synthetic chemistry, digital manufacturing, robotics and fuel development. (U.S. Air Force graphic / Gregory Gerken) Photo Details / Download Hi-Res Machine learning is a subfield of artificial intelligence that gives computers the ability to learn from experience and, eventually, operate without being given explicit instructions, according to Carnegie Mellon University’s School of Computer Science webpage. Systems can adapt by using algorithms and statistical models to analyze and draw inferences from data patterns, information that scientists and engineers can then use to solve problems without having to undergo as many tedious trial-and-error processes. The winning team, led by Carnegie Mellon University Associate Professor Dr. Olexandr Isayev and University of North Carolina at Chapel Hill’s Associate Professor of Chemistry Dr. Frank Leibfarth, proposed a solution for a machine learning-artificial intelligence system that has potential applications for synthetic chemistry, digital manufacturing, robotics and fuel developments, among others, Baldwin said. As the Phase I winner, the team was awarded approximately 30% of a potential $500,000 contract to develop and deliver a tangible solution. The remainder of the contract funds will be awarded incrementally across three additional distinct development phases over the course of nine months, pending approval at each stage. “The point of a Grand Challenge is to put a large enough pot of money out there as an incentive for small businesses, startups and teams in academia to bring their innovative approaches to bear to help solve a problem for the Air Force and Space Force,” said Sean Mahoney, chief intrapreneur, AFRL’s Small Business Office. “In this rapidly changing landscape, it is very difficult for us to keep tabs on the state-of-the-art and what new innovative approaches are bringing breakthroughs that could benefit our warfighters. Grand Challenges are an excellent model to bring those exciting new developments into the labs.” A Congressionally authorized program office for the DOD, National Security Innovation Network, or NSIN, published an initial call for Grand Challenge proposals, also referred to as white papers, on AFRL’s behalf beginning in late September 2022. From there, AFRL selected a small handful of initial white paper submissions and offered 30-minute time slots to each team to officially pitch their proposal in a presentation before they selected the winning team. “UNC brings incredible flow chemistry and synthetic chemistry expertise, whereas Carnegie Mellon’s domain expertise is building algorithms for reinforcement learning and developing digital and AI tools,” Baldwin said. “They complement each other’s skill sets well and, as a result, can tackle some very intriguing problems.” This year’s Grand Challenge, according to NSIN’s website, stems from the fact that the existing body of research evolves so quickly that researchers can struggle to keep up with the absorption of available information. Thus, challenge participants were encouraged to develop a machine learning-artificial intelligence framework to allow researchers to solve problems and answer questions much more quickly, saving time and energy and giving the U.S. military a competitive edge. “As we become more skilled at controlling processes, making new manufacturing processes and gaining access to new synthetic chemistry routes, this means that there is more that we can fine-tune, more processes that we can change for the better,” said Dr. James Hardin, materials research engineer, AFRL’s Materials and Manufacturing Directorate. “But as the number of things that can change increases, it becomes untenable to keep up — even with a fleet of researchers tackling a specific problem all at once.” In order to get more research value out of fewer human hours, researchers can use artificially intelligent machines to comb through multiple possibilities and smartly suggest the next best solution to try, ultimately saving time and money for the U.S. Department of the Air Force, Hardin said. “Even though we may start in relatively confined areas — be it 3D printing or synthetic chemistry — what we learn in those spaces allows us to start scaling it to more application-specific areas,” Hardin explained. “It can potentially lead to things like batteries with longer life, greater power for aircraft. New skins for airplanes, airframes that are lighter, stronger and stiffer, embedded sensing technologies. It gets us closer to that ‘holy grail’ of being able to optimize across both form and function.” About AFRL The Air Force Research Laboratory, or AFRL, is the primary scientific research and development center for the Department of the Air Force. AFRL plays an integral role in leading the discovery, development and integration of affordable warfighting technologies for our air, space and cyberspace force. With a workforce of more than 11,500 across nine technology areas and 40 other operations across the globe, AFRL provides a diverse portfolio of science and technology ranging from fundamental to advanced research and technology development. For more information, visit www.afresearchlab.com.