New Directions at DARPA

By Ed Lazowska and Dave Patterson

In July, Dr. Regina Dugan was sworn in as the 19th Director of the Defense Advanced Research Projects Agency. In September she visited six universities: Virginia Tech (her undergraduate alma mater), Texas A&M, UCLA, Caltech (her graduate alma mater), Stanford, and UC Berkeley. Here are some of the messages she delivered during these visits:

• DARPA recognizes that it must work hard to win back the attention of top researchers in important fields, including computer science.
• DARPA places a high value on fundamental research, framed within its agency mission.
• DARPA recognizes that breakthrough work is not usually characterized by improvements that are monotonic, step-by-step, inch-by-inch. Program managers will work with PI’s to develop the most effective and sensible ways to assess progress. Go/no-go decisions on strict schedules will not be the norm.
• Program managers will be able to have far greater collaboration with their PI’s because they are not in the role of “annual evaluator.” PM’s will be challenged to become active partners with PI’s.
• DARPA is challenging itself, from this point forward, to eliminate, wherever possible, classification, pre-publication review, ITAR and foreign national restrictions, as well as other perceived or actual constraints on fundamental research, except in those unusual cases where it is necessary to protect national security or the safety of military personnel.
• There will be more realistic conflict-of-interest rules for individuals working at DARPA, making it easier for university researchers to spend time at the agency. There will be less adverse impact on the individual, and on his or her home institution.
• Additionally, DARPA will work in partnership with universities and the private sector to ease the transition back to university life after a tour at DARPA
• There will be a renewed emphasis on students—on their ability to surprise us with outside-the-box thinking. (Landmark student achievements in computer science and other fields were highlighted in Dr. Dugan’s presentations.)
• Each and every DARPA program need not have direct, immediate impact on the military. Some will, but others will need to be embraced by industry, or will need further research. Asking the military to try to apply programs in these latter categories is not helpful to these services or to the research community.
• There must be active dialogue between DARPA and universities to develop a sense of shared responsibility. DARPA depends on universities to encourage their best people to serve the nation by working on problems of critical importance, and by contributing to the program management function, which is critical to the quality of research sponsorship.

These messages signal a dramatic shift from the approach of Dr. Dugan's predecessor‐a shift that is of enormous importance to our field and to the nation. Another important signal is the recruiting of our colleague Peter Lee from CMU to DARPA to direct a new Transformational Convergence Technology Office, focused on potentially game-changing fundamental research involving universities and other broader research communities, some well outside the DoD world view (more on this in a later issue). These appointments were preceded by the selection of Zach Lemnios, Chief Technology Officer at MIT Lincoln Laboratory, as the Director of Defense Research & Engineering (essentially, DARPA's "boss").

It is now up to each of us, and each of our institutions, to seize the opportunity and re-engage.

Ed Lazowska holds the Bill & Melinda Gates Chair in Computer Science & Engineering at the University of Washington. 
Dave Patterson holds the E.H. and M.E. Pardee Chair of Computer Science at UC Berkeley. 
Both are past Chairs of the Computing Research Association.