By James Crowley
Budgets, especially the President’s budget for an upcoming fiscal year, always reflect a bit of wishful thinking. The President’s budget is only the first stage in the process of allocating federal funds to programs. Both houses of Congress will have their say, and with any luck an appropriations bill will emerge some time near the October 1 beginning of the fiscal year. In a less auspicious year (often, as in the current case, an election year), an appropriation is delayed or even replaced by a continuing resolution that limits spending to some fraction of the previous year’s.
Nevertheless, budgets reflect the research priorities of the agencies and the administration, and are an indication of the major themes likely to appear in the coming year’s programs.
The Obama administration released its budget request for fiscal year 2013 in early March 2012. Embedded in the budget is information from the federal research-funding agencies (including the National Science Foundation, the Department of Energy’s Office of Science, and the several Department of Defense agencies that fund science) about their plans for fiscal year 2013, pending appropriation of funds by Congress.
Here is what can be gleaned about the proposed budgets for some of the programs of interest to the SIAM community.
Applied mathematics and computational science are part of a number of programs across the broad spectrum of research funded by NSF. The Division of Mathematical Sciences is the home of 11 core programs, including one in applied mathematics and another in computational mathematics, as well as several important interdisciplinary programs.
Like any agency, DMS must respond to national priorities, as set by the administration, if its budget is to grow. A budget that fails to grow is able to do less over time, as the organization’s buying power decreases; ultimately, the result is funding for fewer researchers and students. All the U.S. federal science agencies share some current priorities: manufacturing, innovation, cybersecurity, education, and “big data.” Several agencies plan initiatives in big data: ways to handle massive sets of data and to extract useful information from them (see sidebar). National priorities have played a bigger role than usual this year in defining new research initiatives.
In 2012, despite a 2.5% increase in the overall NSF budget, DMS saw a small decrease, due in part to an NSF-wide shift of money from research to large facilities. DMS director Sastry Pantula has spoken in recent interviews of his desire to protect the core programs. Nevertheless, it is apparent from the range of new programs listed below that additional funding often comes through the creation of interdisciplinary programs.
Funding for CISE, NSF’s Computer & Information Science & Engineering Directorate, grows by 8.5% in the 2013 budget, because computer science is perceived as essential to addressing several key priorities, including big data and cybersecurity. As Farnum Jahanian, NSF assistant director for CISE, reported to the SIAM science policy committee at its April meeting, “Think of basic research, but also how it fits into societal priorities.”
Recently developed NSF programs related to priority areas include:
Some of these programs have a home in CISE or OCI, with DMS as a partner. All are interesting to some segment of the SIAM community.
NSF has developed new modes of funding and new procedures for reviewing proposals, along with new programs that cut across traditional areas. New NSF-wide programs include:
The Advanced Scientific Computing Research program, known as ASCR, is the main program in the Office of Science that provides significant funding for research in applied mathematics and scientific computing.
The ASCR budget continues to emphasize multiscale/complex systems and, for 2013, adds data-intensive science. The latter is part of the national effort across several agencies to focus research on the problem of massive data, says ASCR head Dan Hitchcock. The ASCR budget proposes an allocation of $68.5 million for exascale computing, but data-intensive computing is the new piece in the ASCR program.
Hitchcock stresses the need for new algorithms, both compute-intensive and data-intensive, for evolving computer hardware. Because DOE is a mission agency, programs must respond to agency priorities. Accordingly, ASCR’s data-intensive computing initiative focuses on kinds of data relevant to DOE, such as that coming from large facilities. And, Hitchcock notes, “The data-intensive computing initiative should be seen as part of the general push toward exascale.” He points out that problems associated with data involve uncertainty quantification, and that research must also take into account the multiscale/multiphysics nature of the phenomena being investigated.
New to ASCR for 2013 will be large math centers, as described in the recent an-nouncement of funding for Mathematical Multifaceted Integrated Capability Centers (MMICCs). In the past, ASCR’s mathematics program funded about 110 projects at any time, with support mainly for individuals or small teams of five or fewer investigators.
The MMICCs program is not a small effort. As estimated in the call for pre-proposals, a total of $9 million per year would be available. From this total, ASCR anticipates funding three or four centers, at about $2 million to $3.5 million each per year, with an additional nine teams funded at somewhat lower levels (each from $250,000 to $3.5 million per year). It is anticipated that the centers would be funded for five years, after which they could recompete.
The MMICCs are intended to address long-term mathematical problems arising from grand challenges in areas of interest to DOE. Such challenges include accelerating the discovery and design of new materials and chemistry for energy applications; developing methods for modeling and control of complex engineered systems, such as the U.S. power grid; methods for remediation of contaminants in subsurface flow and/or geologic sequestration of carbon dioxide; managing data at large-scale DOE facilities. Each raises a different set of mathematical challenges, from modeling to computational methods, optimization, and control.
Hitchcock describes the drive toward larger centers as motivated by the need for larger teams with diverse skills to tackle the more complex problems. “The trend is to do team-oriented research,” he says, “but we continue to support individual PIs." With mission-driven compute- and data-intensive challenges and significant changes in hardware, DOE applied mathematics needs to do more with the funding available. These centers could signal the start of major changes in the way the applied math program at DOE funds science.
What happens if the appropriations process produces less money than needed for the proposed research programs? Would the new programs and directions be derailed, waiting for future funds? Maybe. But more likely, reduced versions of the programs would be implemented through existing programs. In any event, the themes found in the budget are a roadmap to future federal investments.
Earlier this year the White House announced a major research initiative on challenges presented by extremely large data sets that are difficult to manage and analyze with today’s techniques. NSF, in concert with other agencies, including the National Institutes of Health and the Department of Energy, has released a new solicitation, Core Techniques and Technologies for Advancing Big Data Science and Engineering.
As described in NSF’s 2013 budget request, the initiative will support the development of new tools and approaches for addressing “the challenges of managing, analyzing, visualizing, and extracting useful knowledge from large, diverse, distributed, and heterogeneous data sets. This includes the development of data analytics, algorithms, and statistical and mathematical methods.”
According to the interagency solicitation, “big data” refers not “just to the volume of data, but also to its variety and velocity. Big data includes large, diverse, complex, longitudinal, and/or distributed data sets generated from instruments, sensors, Internet transactions, email, video, click streams, and/or all other digital sources. The focus is on core scientific and technological advances (e.g., in computer science, mathematics, computational science and statistics).”
This program also has a major interdisciplinary component. It is, as described in the solicitation, “one component in a long-term strategy to address national big data challenges, which include advances in core techniques and technologies; big data infrastructure projects . . . and a comprehensive integrative program to support collaborations of multi-disciplinary teams and communities to make advances in the complex grand challenge science, biomedical research, and engineering problems of a computational- and data-intensive world.”