In a recent study sponsored by the Business Roundtable and the nonprofit group Change the Equation, 97 percent of the CEOs of major American companies identified a lack of science, technology, engineering and math (STEM) skills among the national workforce as a problem for their businesses. Over the next five years these firms will need to hire approximately one million new employees with these skills and more than 600,000 with applied science backgrounds. The nation has been in this situation before. In 1944 President Franklin Delano Roosevelt commissioned Vannevar Bush, director of the wartime U.S. Office of Scientific Research and Development, to create a plan for sustaining the momentum of scientific achievement that had occurred during World War II, in part by responding to a looming skills shortage—the result of the large number of potential students who had been drafted into the military.

One of the biggest impacts of Bush's report was to create the separation of basic and applied research, a model that predominates in federally funded science today. Although this separation has been very effective in many fields, in education and the social sciences, basic research sometimes fails to translate successfully into applied settings. As we study ways to confront this new crisis, we should consider an alternative approach to research—one that the Defense Advanced Research Projects Agency has been demonstrating since 1958.

The DARPA process is reminiscent of the development cycles for radar and the atomic bomb during WWII: diverse teams of the brightest minds iterate continuously on basic research challenges aimed at solving enormously complex problems. Unlike traditional basic or applied research, the DARPA method resides in a category that the late science policy researcher Donald E. Stokes introduced in his 1997 book, Pasteur's Quadrant. In Stokes's classification, basic research resides in Bohr's quadrant: it is the quest for basic knowledge without regard for the final use of that knowledge. Applied research lies in Edison's quadrant, where producing a specific product is the top priority. In Pasteur's quadrant, named after Louis Pasteur, basic research is applied to solve specific and immediate problems.

As an agency, DARPA lives in Pasteur's quadrant. Every project is a moon shot. The final goal is clear, but the process for getting there remains flexible. In the U.S. Department of Education's Office of STEM, we have been proposing the use of Pasteur's quadrant as a means for creating so-called moon shots for education, especially at the intersections of science and technology. The possibilities for research are plentiful. Can customized digital tutors be created that adapt to the student over the course of his or her education, from preschool through college? Can these same educational technologies be developed in ways that encourage and enhance lifelong learning? Can we find new approaches to assessment that measure mastery in real time rather than at the end of a course? If solutions such as these are possible, they will be achieved only by bringing together the most innovative teams of researchers, professional developers and educators to tackle the problems as a whole. That is why President Barack Obama's 2016 budget proposes up to $50 million for an Advanced Research Projects Agency–Education (ARPA-ED) to allow the Department of Education to support rapid-cycle, high-impact technology development aimed at preparing students for the 21st-century workforce.

To determine where investment should be made, my colleagues and I are currently convening groups of innovators and educators to evolve a vision of STEM education in 2025. Once that vision is clear, we will deconstruct it and outline a plan for achieving it. Will that vision be the correct one? It is hard to say, but this initial vision does not have to be absolutely correct. As long as the basic target of improving educational outcomes remains in sight, the goal and the vision can be adjusted as we work toward them. Just as DARPA researchers could not have predicted what the Internet would become when they laid its foundation in 1968, today's innovators will not know how technology can transform education until they roll up their sleeves and do it.