As I noted in the last post, the Senate Appropriations Committee included language in the Senate verision of the FY 06 Defense Appropriations bill that would strip $55 million from an element of DARPA’s Cognitive Computing program — a move that seems to run counter to recent congressional sentiment on the role of computer science, especially university-led fundamental computer science, at DARPA. The “out of the blue” cut — there was no advance warning provided by appropriations staffers, no evidence that there was dissatisfaction with the program — would fall on the “Learning, Reasoning, and Integrated Cognitive Systems (pdf)” (COG-2) account. With little feedback from the Senate appropriators, we’re operating under the assumption that the main impetus for the cut was to provide an offset for increases elsewhere in the bill, though there’s been some speculation among Senate staff that the program may have suffered do to a misperception that it’s somehow similar to some of the agency’s more controversial bio-related programs.
It’s not. As we’ve tried to point out to the conferees who will have to determine whether the cut will stand in the compromise bill negotiated with the House (the House opted to fund the program at the agency’s requested level), research in learning, reasoning, and cognitive systems is focused on intelligent intrepretations of signals and data, on controlling unmanned vehicles, and on amplifying human effectiveness. Its aim is to reduce U.S. casualties by providing improved command and control and tactical planning against adversaries, as well as improved training systems. Work in this area includes research responsible for the Command Post of the Future (CPOF) — a software system currently deployed and very widely-used in Iraq to coordinate battle plans and integrate multiple intelligence reports, providing U.S. forces the capability to plan, execute and replan much faster than the enemy’s decision cycle and cited by Secretary Rumsfeld as the major contributor to victory in the first phase of Operation Iraqi Freedom. It’s also cricital to the research and development of autonomous, unmanned vehicles that amplify our warfighting capability while reducing the number of U.S. forces in harm’s way. Cutting support so significantly for this research will hamper advancements in defense-related IT in the short- and long-term and will slow technological advancements essential to current and future military operations in Iraq and around the globe.
It also runs completely counter to recent concerns of Congress, PITAC and the DOD’s Defense Science Board. All three bodies have raised strong concerns about the shift of DARPA resources away from fundamental research at universities, especially in information technology. The Cognitive Computing program is one area where DARPA has responded positively to these concerns.
Because these “out of the blue” cuts are so difficult to counter — they appear very late in the process with very little information about what motivated them and at a time when access to appropriations staff is most limited — we’ve focused our strategy opposing the cut by urging the members of the conference committee to abandon the Senate number and adopt the President’s budget request, the number approved by the House. You can help. We’ll tell you how soon….
Washington Post columnist Sebastian Mallaby has an interesting op-ed today inspired by news of Microsoft chairman Bill Gates’ tour of college campuses, urging students to consider majoring in computer science. The piece does a good job of making the case that Gates makes in his talks to students — computing is a field with a history of producing really great stuff that promises to make even more really great stuff in the future.
In most fields of human endeavor, you hope for gradual improvements: an engine that’s somewhat more efficient, a medicine that improves life expectancy by a few months. But computer power progresses exponentially, warping social life, intellectual horizons and the business playing field.
And Mallaby lays out some examples:
Smart watches will download weather forecasts and news headlines over wireless connections. Smart phones will scan products in department stores to check where better prices can be found. Notebook computers will be portable libraries with the weight of just one novel — libraries that allow you to scribble in the margins and share your witty insights wirelessly with friends. Your home computer will respond to instructions both written and vocal, and it won’t be a computer so much as a network. Music, videos, games, photographs — oh, yes, and all your lofty intellectual outpourings — will be beamed around the house to a variety of screens and speakers. The tablet on the kitchen counter will display recipes and shopping lists. The plasma screen on the wall will be for family photos.
…All great examples of some of the foreseeable future in computing. And Gates deserves enormous credit for taking on this role of cheerleader for the field. With the current trends facing the discipline, and a general trend of US students shying away from careers in math and the sciences, this effort is sorely needed.
The only additional thing I’d wish for in these kind of presentations, especially for prospective students, would be to add more sense of purpose to the call. We’ve had a very interesting discussion about this piece amongst the members of CRA’s Government Affairs Committee, including this great observation about what could be said about the “calling” of a career in science generally, and computer science in particular, from Peter Lee at CMU (who gave me permission to post it here):
Choosing to devote your life to science and technology is not a “normal” or “safe” choice. It is a choice made by people who are exceptionally smart, caring, and idealistic. Science makes people smarter and less scared, and it also makes the world better. Becoming a scientist means joining a community of idealists.
There’s not much of that in the talks that Gates is giving, but that’s understandable. It’s easy to lose sight of intellectual and ideological appeal when the practical applications are so plainly visible.
Anyway, a digression from the piece, but something that occurred to me and many of the other members of the committee.
The meat of Mallaby’s piece comes in the final three paragraphs though, where he’s right on the money:
A lot of Washington debates are about managing bad stuff: war, terrorism, natural disasters, killer viruses, budget deficits, trade deficits, medical inflation, airline bankruptcies, imploding corporate pension plans. But policy also needs to focus on the good stuff: To figure out how we can accelerate progress. If we don’t fix the budget deficit, we will be setting ourselves up for economic punishment. But if we don’t position ourselves to take advantage of technology, we will be setting ourselves up to miss a huge economic prize.
What must we do to remain prize-worthy? The good news is that, in Gates’s estimation, between 17 and 19 of the world’s top 20 computer science faculties are American, and Microsoft hasn’t yet moved many software jobs offshore. But to keep things that way we need to step up federal research funding and relax post-Sept. 11 visa rules, so that the United States remains what Gates calls “an IQ magnet.” And because smart Indians, Chinese and others are more likely to return home as their countries grow freer and more prosperous, the United States must focus on growing its own talent. Last year two respected global surveys of math skills in eighth and ninth grades put the United States in 15th and 24th place, respectively. That isn’t good enough.
It would take fairly little to address these problems. Last week a panel convened by the National Academies proposed a package of measures that ranged from math prizes for high schoolers to pay raises for math teachers, along with a program to boost federal research funding by 10 percent annually for seven years. The total price tag comes to $10 billion annually, but the nation spends nearly twice that amount on absurd farm subsidies. What kind of priorities are those?
Maybe Mallaby’s seen that argument somewhere else 
The bad news is that the stars are aligning in such a way as to guarantee that there will be no increase for computer science, or the sciences generally, in the foreseeable future. The Republican Leadership is being pushed hard at the moment to find funds to “pay for” the large emergency supplements paid out for hurricane relief. Odds are those funds will come through across-the-board cuts to non-defense, non-security related discretionary spending. Look for science agencies to suffer cuts similar to last year’s across-the-board 2 percent reduction (or worse).
One particular computing program is under an even bigger threat. The Senate voted to approve a $55 million cut to DARPA’s cognitive computing program as part of the FY 06 Defense Appopriations bill. The out-of-the blue cut would hit DARPA’s $114 million “Learning, Reasoning, and Integrated Cognitive Systems” account, effectively cutting the program in half. The House did not call for a cut in its version of the bill, so CRA is working to urge members of the conference committee to abandon the Senate cut and embrace the House number. We’ll have all the details in the next post. I just thought it worthy of mention that at the same time the calls keep coming for increased support of computer science and the physical sciences, and as much progress as has been made to draw the linkage between federal investment in university research and our ability to continue to innovate, a significant percentage of our policy leadership still doesn’t get it.
This by now has been covered all over the place, but I’d be remiss not to add it here, too. The National Academies convened a 20-member panel last summer at the request of Sens. Lamar Alexander (R-TN) and Jeff Bingaman (D-NM), and Reps. Sherwood Boehlert (R-NY) and Bart Gordon (D-TN) to determine the “top 10 actions, in priority order, that federal policymakers could take to enhance the science and technology enterprise so the United States can successfully compete, prosper, and be secure in the global community of the twenty-first century.” The task was fast-tracked, and an august panel was put together, chaired by former Lockheed Martin CEO Norm Augustine and including folks like Intel Chairman Craig Barrett, President of Texas A&M Robert Gates, CEO of DuPont Charles Holliday, Jr., former Director of Defense Research and Engineering at DOD (and computer scientist) Anita Jones, and MIT president emeritus Chuck Vest. The committee met once, held focus groups on the five issue areas they decided merited attention (K-12 education, higher education, research, innovation and workforce issues, and national and homeland security), then put together the report they released on Wednesday, Rising Above the Gathering Storm: Energizing and Employing America for a Brighter Economic Future.
The committee actually came up with four major recommendations, supplemented with 20 specific actions to be taken. Hard to quibble with the broad recommendations:
Recommendation A: Increase America’s talent pool by vastly improving K-12 mathematics and science education.
- Recruit ten thousand teachers, educate ten million minds — a program that would award competitive 4-year scholarships for students to obtain bachelor’s degrees in the physical or life sciences, engineering or mathematics with concurrent certification as K-12 math and science teachers.
- Strengthen two hundred fifty thousand teachers’ skills, inspire students every day — provide summer institutes, science and mathematics master’s programs, advanced placement training, and a curriculum modeled on world-class standards for current K-12 teachers.
- Enlarge the pipeline by creating opportunities and incentives for middle-school and high-school students to pursue advanced work in science and math.
Recommendation B: Sustain and strengthen the nation’s traditional commitment to the long-term basic research that has the potential to be transformational to maintain the flow of new ideas that fuel the economy, provide security, and enhance the quality of life.
- Increase the federal investment in long-term basic research by 10 percent annually over the next 7 years, with special attention paid to the physical sciences, engineering, mathematics, and information sciences and to DOD basic-research funding.
- Provide new research grants of $500,000 each annually, payable over 5 years, to 2000 of our most outstanding early career researchers.
- Institute a National Coordination Office for Research Infrastructure to manage a centralized research-infrastructure fund of $500 million per year over the next 5 years.
- Allocate at least 8 percent of the budgets of federal research agencies to high-risk, high payoff research.
- Create in DOE an organziation like DARPA [hopefully in the model of "old" DARPA - ed] called ARPA-E which would be charged with R&D to meet the nation’s long-term energy challenges.
- Institute a Presidential Innovation Award to stimulate scientific and engineering advances in the national interest.
Recommendation C: Make the US the most attractive setting in which to study, perform research, and commercialize technologic innovation so that we can develop, recruit, and retain the best and brightest students, scientists and engineers from within the US and throughout the world.
- Provide 25,000 new 4-year undergraduate scholarships each year to US citizens attending US institutions.
- Increase the number of US citizens pursuing graduate study “in areas of national need” by funding 5,000 new graduate fellowships each year.
- Provide a federal tax credit to encourage employers to make continuing education available to practicing scientists and engineers.
- Continue to improve visa processing for international students.
- Provide a 1-year automatic visa extension to international students who receive doctorates or equivalent in STEM or other areas of national need at US institutions to remain in the US to seek employment.
- Institute a new skills-based, preferential immigration option.
- Reform the current system of “deemed exports”.
Recommendation D: Ensure that the US is the premier place in the world to innovate, invest in downstream activities, and create high-paying jobs that are based on innovation by modernizing the patent system, realigning tax policies to encourage innovation, and ensuring affordable broadband access.
- “Enhance” intellectual-property protection for the 21st century global economy.
- Enact a stronger R&D tax credit to encourage private investment in innovation.
- Provide tax incentives for US-based innovation.
- Ensure ubiquitous broadband Internet access.
Hard to find fault in much of that — though I’m leeriest of the IP-related “enhancements” (see the report for the details about each of the action items listed). The committee came up with a “back of the envelope” calculation of about $10 billion annually to fully implement the recommendations (the R&D tax credit recommendation is actually the costliest). While that number might seem impossibly high to achieve under the current political mindset for science funding — after all, NSF suffered a 2 percent cut in the last budget and the smart money is betting something similar for FY 06 when appropriations finally wrap up — in the grand scheme of things, $10 billion on top of an $840 billion discretionary budget is a relatively small investment for the potential benefit. If the President is looking for an initiative that would enhance his legacy, I think he’d be hard-pressed to find one with a better cost/benefit ratio.
Anyway, as I said, the report has gathered a reasonable amount of attention in the press. The New York Times has coverage, as does the Chronicle of Higher Education (subscription required). There are a couple of follow-up hearings planned, including one by the Senate Committee on Energy and Natural Resources on October 18th, and the House Science Committee on October 20th.
We’ll try and have all the details here.
(Thanks to Sam Liles and Spaf for the pointers!)
From the CCC Blog
- DARPA Issues PERFECT Solicitation Erwin Gianchandani
- Microsoft Announces New Policy-Focused Internship Program Erwin Gianchandani
- NAE Elects its Class of 2012 Erwin Gianchandani
- NIST to Fund Pilot Projects Advancing Trusted Identities Erwin Gianchandani