PCAST Finds IT R&D Critical to U.S. Competitiveness, Calls for Renewed Federal Investment

By CRA

Backed by strong support from the White House, the President’s Council of Advisors on Science and Technology (PCAST) released a biennial report about the status and direction of the nation’s 14-agency, $4.3 billion Networking and Information Technology Research and Development (NITRD) program last month. The report emphasized the critical role of advances in networking and information technology (NIT) to U.S. economic competitiveness, and called on the nation to “continue to innovate more rapidly and creatively than other countries in important areas of NIT” in order to sustain and improve overall quality of life.

The nation’s Chief Technology Officer, Aneesh Chopra, and Chief Information Officer, Vivek Kundra, joined other leading White House officials, co-chairs of PCAST’s NITRD review working group, and external discussants at the public release of the “Designing a Digital Future” report in Washington, DC on December 16.

Tom Kalil, White House Office of Science and Technology Policy Deputy Director for Policy, one of the participants in the report’s roll-out, summarized why NIT R&D is so important to the current administration:

• The information revolution is far from over, with a number of core challenges in computer science yet to be addressed;
• NIT can accelerate the pace of discovery in virtually every scientific domain;
• NIT is having a huge impact on the nation’s economy, with over two-thirds of the increase in productivity since 1995 due to NIT; and
• There exists a very direct relationship between NIT and our nation’s grand challenges—healthcare, education, energy, and open government, among others.

Kalil’s comments were consistent with PCAST’s key findings about the status and direction of NIT R&D:

“Advances in NIT are crucial to achieving our major national and global priorities...  NIT will be an indispensable element in buildings that manage their own energy usage; attention-gripping, personalized methods that reinforce classroom lessons; continuous unobtrusive assistance for people with physical and mental disabilities; and strong resilience in cyberwarfare.” Ultimately, “NIT advances accelerate the pace of discovery in nearly all other fields.”

And they came just days after Larry Summers, then the outgoing Director of the National Economic Council, delivered impassioned remarks about the significance of information technology to the nation’s economy at the unveiling of another PCAST report (details about that event and report):

“If you look at the economic history of the last 150 years, it’s a lot about the steam engine, it’s a lot about electricity, it’s a lot about things that came from the automobile—all of which gave people capacities to do things they hadn’t done before and touched almost every aspect of life. And this generation’s technology is information technology with all that it makes possible, and we need to make sure that it is exploited as widely and as pervasively as possible. And what better time to accelerate investment in information technology than at a time of substantially unemployed resources that can be put to work providing jobs that are important in the short term and providing capabilities that are profoundly important in the long run.”

Key Recommendations by PCAST

PCAST makes three key recommendations in the “Designing a Digital Future” report, beginning with a call for new multi-agency NIT R&D initiatives that support high-risk/high-reward research in areas of particular importance to national priorities:

• Health information technology:  make possible comprehensive lifelong multi-source health records for individuals; enable both professionals and the public to obtain and act on health knowledge from diverse and varied sources as part of an interoperable health IT ecosystem; and provide appropriate information, tools, and assistive technologies that empower individuals to take charge of their own health and healthcare to reduce its cost. Importantly, PCAST states in the report that we must “[go] well beyond the current national program to adopt electronic health records.”
• Energy and transportation:  dynamic power management broadly; interoperable standards for real-time control; low-power systems and devices; and improved surface and air transportation.
• Security and robustness of cyber-infrastructure:  more effective ways to build trustworthy computing and communications systems; better defense mechanisms for today’s infrastructure; and fundamentally new approaches for making cyber-infrastructure truly resilient to cyber-attack, natural disaster, and inadvertent failure.

PCAST further calls for increased investment in a number of fundamental NIT research areas that will accelerate progress across a broad range of priorities, including:  privacy and security; human-computer interaction (“human-machine interaction and social collaboration and problem-solving in a networked, online environment”); data analytics, including data collection, storage, management, and automated large-scale analysis based on machine learning and predictive modeling; and computing in the physical world through advanced sensor and control networks, innovative robotics, etc.

A third PCAST recommendation is to create a new standing committee of leading academic and industrial NIT experts to help guide NITRD and provide strategic advice for the program.

Improving How We Classify NIT R&D

As part of its NITRD review, PCAST found that current budget reporting mechanisms do not accurately detail the Federal investment in NIT R&D. Instead, many agencies use “today’s information technology” in order to advance R&D in their own fields, leading to substantial overstatement of the true Federal investment in core NIT R&D. For example, a PCAST-requested review of National Institutes of Health (NIH) commitments in FY 2009 found between only two and 11 percent of the $1 billion reported to NITRD as NIH’s NIT R&D portfolio that year could actually be ascribed as core NIT R&D. The rest corresponded to applications constituting important infrastructure to support R&D in biomedicine and the health sciences; while they are legitimately categorized as R&D expenditures, they are not NIT R&D as they involve using today’s information technology to advance the forefront of other fields, not to drive the forefront of NIT.

As PCAST working group co-chair, Ed Lazowska, of the University of Washington commented: “[This] is an absolutely appropriate expenditure, and it’s an essential expenditure; it’s important that NIT drives forward all other fields of discovery.  On the other hand, these are for the most part not investments that are pushing the forefront of NIT…  What we advocate is simply that the budget categories make this clear, so that the nation knows how much it is actually investing in driving this critical field forward. Today, we believe we are investing far less in advancing this field than one would believe from the budget reports.”

The Race for FLOPS

Another recommendation—in light of the recent news of a Chinese supercomputer heading the list of the world’s fastest machines—is the panel’s call for the U.S. to abandon the competition to stay atop rankings based on floating point operations per second (FLOPS). Describing it as “an arms race we don’t really find beneficial,” PCAST member and working group co-chair, David Shaw, and Lazowska noted that finishing at the top of the FLOPS race is exceptionally expensive and does not guarantee a machine that is particularly useful. A far more valuable priority, Shaw noted, is to invest in research that could allow for a leapfrog of current high-performance computing technology.  “We need to make sure we don’t allow procurement to crowd out research funding,” Shaw said.

Federal Investment Is Critical

Importantly, while industry continues to make major contributions to the field, PCAST notes in its report that we must not “equate the very large industry R&D investment in NIT with fundamental research of the kind that is carried out in universities and a small number of industrial research labs. The vast majority of industry R&D in NIT is focused on development… fundamental research with the potential for future transformational application represents a fraction of overall industry R&D in NIT… [and thus] Federal investment in NIT R&D is and will remain essential.”

Note: See full report at: http://www.whitehouse.gov/sites/default/files/microsites/ostp/pcast-nitrd-report-2010.pdf. Archived webcast of the report’s rollout at: http://lazowska.cs.washington.edu/nitrd/video/. Continuing coverage of the report’s impact will be available in the coming months on the CCC Blog and CRA’s Policy Blog.