Today the world faces extraordinary opportunities enabled by technological progress: new technologies for communication and deep collaboration in teams and new ways to collect and process information that are transforming industries and computing technologies and offer the promise of higher levels of automation and intelligence in the devices created. Today’s engineers stand on the cusp of dramatic advances in materials, information, robotics, energy, transportation, manufacturing, and health. Use of information technology-enabled tools for collaboration is now the norm, including artificial intelligence and “big data,” allowing teams—both large and small—to work more effectively.
The world also faces a complex set of global challenges: threats to the environment, threats to national security, disruptive changes in the workforce, new diseases and health risks, and a rapidly changing world economy and competitive landscape.
Solutions will require research teams capable of bringing together expertise from both natural and social science disciplines to address all facets of these challenges. This approach has been conceptualized as a continuum of increasing disciplinary integration, starting with multidisciplinary, and proceeding through interdisciplinary to transdisciplinary.1
“Multidisciplinary research is typically understood as the sequential or additive combination of ideas or methods drawn from two or more disciplines or fields to address the focal problem. Interdisciplinary research involves the integration of perspectives, concepts, theories, and methods from two or more disciplines or fields to address the problem. Transdisciplinary research entails not only the integration of discipline-specific approaches, but also the extension of these approaches to generate fundamentally new conceptual frameworks, hypotheses, theories, models, and methodological applications that transcend their disciplinary origins.”2
1 K.L. Hall, A.L Vogel, B.A Stipelman, D. Stokols, G. Morgan, and S. Gehlert, 2012, A four-phase model of transdisciplinary team-based research: Goals, team processes, and strategies, Translational Behavioral Medicine 2(4):415-430.
3 National Research Council (NRC), 2014, Convergence: Facilitating Transdisciplinary Integration of Life Sciences, Physical Sciences, Engineering, and Beyond, The National Academies Press, Washington, D.C.
While convergence of technical science and engineering disciplines is often observed in current research programs, it is less common to see the successful integration of these with the social sciences. Such a comprehensive approach will be important in addressing major societal problems.
FINDING 2-1: This is a time of enormous opportunity in which exponentially expanding knowledge in previously distinct fields can now be combined in new ways to create innovations of great value for society.
The unprecedented growth and convergence of many technologies in recent decades offers the opportunity to address big, grand-challenge-like problems over a time horizon that is within the timeframe of a major National Science Foundation (NSF) initiative. Examples might include the National Academy of Engineering’s (NAE’s) Grand Challenges (e.g., provide access to clean water, Box 2.2). This big-problem focus changes the narrative about engineering and its importance to society, and it dovetails well with an NAE initiative to change how engineering is represented.4
Developing new technologies with high societal impact might be another theme of future centers. One source for such technologies could be the recently identified six “research big ideas” and the three “process ideas”
4 National Academy of Engineering, 2008, Changing the Conversation: Messages for Improving Public Understanding of Engineering, The National Academies Press, Washington, D.C.
identified by NSF (see Box 2.3). The committee considered three of these big ideas—harnessing data for 21st century science and engineering, shaping the new human-technology frontier, and quantum leap: leading the next quantum revolution—as being particularly relevant for future centers, although all six could incubate technologies essential to the solution of problems with great societal impact.
There are many other opportunities and guiding themes for complex engineering and societal problems that could be appropriate for future NSF engineering centers. Examples include the 15 global challenges identified by the Millennium Project5 or the health and development grand challenges identified by the Bill and Melinda Gates Foundation.6 Other entities, such as the President’s Council of Advisors on Science and Technology (PCAST) and the Office of Science and Technology Policy (OSTP), have described similar initiatives, including the National Nanotechnology Initiative,7 the National Robotics Initiative,8 the Brain Initiative,9 and the Cancer Moonshot.10
5 The Millennium Project, “Challenges,” http://millennium-project.org/millennium/challenges.html, accessed September 23, 2016.
8 National Science Foundation, “Funding: National Robotics Initiative 2.0: Ubiquitous Collaborative Robots (NRI-2.0),” https://www.nsf.gov/funding/pgm_summ.jsp?pims_id=503641, accessed October 10, 2016.
10 National Institutes of Health, “Cancer Moonshot,” https://www.cancer.gov/research/key-initiatives/moonshot-cancer-initiative, accessed September 9, 2016.
There appears to be an emerging consensus among international centers that future university-industry research centers should be more challenge-focused—that is, a greater fraction of centers addressing “needs pull” challenges, rather than just tackling “science push” opportunities.11
The “big problem” orientation described above should help to inspire future center personnel, but it complicates the tasks of assembling the right research team, managing it efficiently, and maintaining its focus on center goals. These challenges highlight the importance of the systematic use of the best practices of team research and value creation in the centers to give them the best opportunity to succeed.
Team research12 refers to research conducted by more than one individual in an interdependent fashion, including research conducted by small teams and larger groups. As defined here, it includes all of engineering as well as traditional physical and social science fields. To reach their goal successfully, multidisciplinary research teams must overcome a number of challenges, including the integration of members with different areas of expertise,
11 E. O’Sullivan, 2016, “A Review of International Approaches to Center-Based, Multidisciplinary Engineering Research,” a commissioned paper for this study, available at https://www.nae.edu/Projects/147474.aspx.
12 In this report, which is about engineering research, the committee chose to use the phrase “team research” rather than adopt the earlier NRC report’s phrase “team science,” but the principles are the same.
different vocabularies and ways of approaching problems, different understanding of the problems to be addressed, and different working styles.
There is a strong body of research conducted over decades on how team processes influence team effectiveness.13 While much of the research has been on nontechnical teams, the insights can be applied to engineering research teams. Examples include the following:
- Team composition influences team effectiveness; in particular, task-relevant diversity is critical and has a positive influence on team effectiveness;
- Team professional development training improves team processes and outcomes; and
- Geographically dispersed science teams and groups face more challenges in communicating and developing trust than do face-to-face teams and groups.
Best practices of team research include those listed in Box 2.4. Implementation of these practices is not easy, and requires considerable time.
Adherence to these team research best practices will be particularly important in addressing the challenges of integrating the social science members with the physical science and engineering members of the team.
Value creation is the name for the learning and creating activity whose goal is the development of new, sustainable value for society, whether as notable new research results or as marketplace innovations. As discussed in Chapter 1, top professionals and enterprises today have the innovative skills and value-creation processes to identify and systematically develop major new opportunities. These programs are showing that large systemic improvements in productivity can be made.14 In Box 2.5, the committee lists what in its view are examples of value-creation best practices. Suggested best practices for an enhanced proposal process are provided in Box 6.1.
13 NRC, 2015, Enhancing the Effectiveness of Team Science, The National Academies Press, Washington, D.C.
14 NRC, 2015, Making Value for America: Embracing the Future of Manufacturing, Technology, and Work, The National Academies Press, Washington, D.C.
Most of the systems mentioned in Box 2.5 are oriented toward delivering economic value for industry—the more commonly accepted definition of value creation—but the principles can easily be adapted to the broader goal of delivering societal value.
According to NSF, “the goal of the ERC program is to integrate engineering research and education with technological innovation to transform national prosperity, health, and security.”15 The idea that the Engineering Research Center (ERC) program would provide economic benefits through engagement with industry to enhance U.S. industrial competitiveness and to develop the “innovation ecosystem” has been a common expectation from its inception.
15 NSF, “Engineering Research Centers (ERC),” https://www.nsf.gov/funding/pgm_summ.jsp?pims_id=5502, accessed October 10, 2016.
As the ERC program has evolved, funded centers have tended to be targeted fairly narrowly on specific technology areas (e.g., the ERC for Translational Applications of Nanoscale Multiferroic Systems or the ERC for Revolutionizing Metallic Biomaterials).16 This technology focus has the advantage of simplifying the challenges of determining the disciplines needed to address the problem and identifying the key academic and industry research participants.
The successes of the ERC program are described in Chapter 1. However, interviews and reviews of the performance of engineering research centers in different countries suggest that few realize their full potential. Center researchers have tended to work as individual contributors and thus fail to address the center’s major opportunity.17 The committee believes that implementing the practices in Boxes 2.4 and 2.5 would make a profound improvement in center outcomes.
These centers have value only to the degree that their societal or market impact is larger than their transaction costs. Today, those transaction costs are very high and output is limited: opportunity identification, team selection and alignment, collaboration, and bureaucratic overhead are all significant problems. Additional barriers are the inability to access the most current, relevant knowledge, the inability to access the best new talent when needed, and the inability to rapidly pivot the initiative as required by better real-time understanding of the technology and marketplace.
NSF funding of ERCs, which amounts to between $3 million and $5 million per center per year, was significant in the mid-1980s, but has declined in real value by almost a factor of three since then due to inflation. Today there are multiple funding opportunities that are in this same range available to individual principal investigators (PIs), and there are significant opportunity costs associated with participating in a center. At the same time, the centers perceive that annual reporting requirements and bureaucratic oversight have increased.18
Here and in later chapters the committee articulates a strategic new direction for future engineering centers. This new direction has two components: a “what” and a “how.” The “what” is a shift from the current focus on developing a promising new technology area to addressing a high-impact societal or technological need. The “how” is the systematic use of team-research and value-creation best practices to focus the effort and stimulate innovation.
In the context of engineering, the committee defines the phrase convergent engineering as a deeply collaborative, team-based engineering approach for defining and solving important and complex societal problems. All necessary disciplines, skills, and capabilities are brought together to address a specific research opportunity. It is distinguished by resolutely using team-research and value-creation best practices to rapidly and efficiently integrate the unique contributions of individual members and develop valuable and innovative solutions for society.
RECOMMENDATION 2-1: The National Science Foundation should re-invigorate the engineering research center concept by addressing grand-challenge-like problems whose solutions offer the greatest benefits for society and by adhering to the use of team-research and value-creation best practices, fewer administrative burdens, and greater investment and prestige to attract the superb, diverse talent required.
To emphasize the ambition and the new direction of these center-scale investments led by engineering, they should be given a new name, possibly convergent engineering research centers (CERCs). CERCs should continue the early-stage research focus of current ERCs (fundamental research to proof of concept, or technology readiness level [TRL] 1-3), since higher TRL research is difficult to address in a university environment. However, convergent engineering research is expected to usher in breakthroughs that return lasting societal benefit, create high-technology jobs, and serve as a catalyst to the U.S. innovation economy and security needs. Further, the
18 NSF, ERC Key Features: Designing the Next-Generation ERC, Report from the 2007 Annual Meeting, Arlington, Va.
possibility of solving important societal problems through the power of engineering will be a great attractor of talented undergraduate and graduate students, particularly from traditionally underrepresented domestic groups (see Chapter 3). New partnership models based on the best practices of team research and value creation will help to bridge the academe–industry–other19 divide and will provide rapid and significant translation of theory into new commercial products, services, and industries.
Grand challenges such as those in Box 2.2 from the NAE address complex problems, the solutions of which encompass an array of scientific and engineering disciplines. A single CERC may not suffice to address the totality of the challenges. Rather, a number of CERCs in partnership with other related research centers will likely be needed to address them.
The ongoing Internet and information technology (IT) revolutions are redefining organizations, with much of the value coming from transactional efficiency—removing barriers, middlemen, delays, facilities, logistics, etc. There are many examples: autonomous vehicles, ride and residence sharing, real-time 3-D printing, virtual-personal assistants, online shopping, and so on.
The NSF CERC of the future will be part of this transformation. Thanks to high-speed Internet connectivity, engineering and science research and education have become increasingly collaborative efforts on a global scale, leading to the creation of engineering and science platforms accessible to diverse collections of colleagues around the world. These platforms can be physical (large-scale systems, such as the Large Hadron Collider at CERN, or shared manufacturing test facilities), virtual (shared software systems, such as those hosted on GitHub20), or data oriented (shared data collection sites, such as GalaxyZoo21 and OpenStreetMap22). CERCs will very likely employ such creation and collaboration platforms, and future centers may be built around them rather than around physical locations.
The rise of data science is one of the biggest changes since the conception of the ERCs in the 1980s. Data science employs techniques and theories drawn from many fields within the broad areas of mathematics, statistics, operations research, information science, engineering, and computer science.23 Methods that scale to big data are of particular interest in data science, although the discipline is not generally considered to be restricted to such big data, and some big data technologies are focused on organizing and preprocessing the data instead of analysis. The development of machine learning has enhanced the growth and importance of data science and analysis for a variety of research problems.
In regards to CERCs, breakthroughs will be powered by advanced computing capabilities that help researchers manipulate and explore massive data sets. The speed at which any given scientific discovery advances will depend on how well its researchers collaborate with one another, and with technologists, in areas of e-science such as databases, workflow management, visualization, and cloud-computing technologies.24 It is anticipated that the application of data science methods and approaches will affect advances in many fields of engineering, including machine translation, speech recognition, robotics, search engines, the digital economy, as well as the biological sciences, medical informatics, health care, social sciences, and the humanities. From an engineering perspective, data science has fostered competitive intelligence, a newly emerging field that encompasses a number of activities, such as data mining and data analysis.25,26
CERCs could uniquely benefit from incorporating methods and approaches from this emerging field to help further gain insights about trends and patterns in data that will foster research breakthroughs. It will be beneficial for particular CERC teams to leverage advances in data science and analytics and to recruit data science experts to join their research teams.
19 Others include nonprofit organizations, government laboratories, etc.
23 J. Foreman, 2013, Data Smart: Using Data Science to Transform Information Into Insight, Wiley & Sons, p. xiv.
24 S. Tansley and K.M. Tolle, 2009, The Fourth Paradigm: Data-Intensive Scientific Discovery, Microsoft Research, Redmond, Wash.
25 M. LaPonsie, 2011, “Data Scientists: The Hottest Job You Haven’t Heard of,” Onlinedegrees.com, August 10, https://www.aol.com/article/2011/08/10/data-scientist-the-hottest-job-you-havent-heard-of/20007479/.
26 T. Nguyen, 2015, “Data Scientists vs Data Analysts: Why the Distinction Matters,” import.io, October, https://www.import.io/post/datascientists-vs-data-analysts-why-the-distinction-matters/.
Top-Down Versus Bottom-Up Approach to Identification of Research Topics
Mission agencies such as the Defense Advanced Research Projects Agency (DARPA) or the Department of Energy (DOE) typically use a “top-down” method of setting research priorities; that is, the funding agency puts out a thematic call for proposals in a general area of research or with a particular problem in mind.27 In contrast, the NSF typically relies on research proposals developed on topics of particular interest to the proposers and submitted to the funding agency (“bottom-up” approach), although there have been times when NSF requested proposals in some targeted areas, such as manufacturing or biotechnology. As one example of the bottom-up approach, Stanford University’s School of Engineering conducted an exercise in 2015 in which faculty, staff, and students were asked to reach consensus on important, strategic directions for engineering research over the next 20 years.28 In this approach, centers would have considerable discretion in defining their research direction, as long as they meet NSF’s overall objectives.
Compared with top-down methods, bottom-up approaches may facilitate stronger ties with local research ecosystems, involving academic institutions, established corporations, and venture capital communities. Such ecosystems may also provide a more diverse pool of talent for centers. CERCs can be vehicles to enable transformative research or solve challenges of a local or regional nature. (See the example of the federal-state-local partnership model outlined in Chapter 7.)
Top-down approaches focus on well-recognized problems that are agency priorities, while bottom-up approaches can tap the genius of a broad cross-section of the research community and may promote local and regional economic development. Both approaches help to attract and retain a diverse cadre of researchers, and a mixed approach may have merit for CERCs.
Because the proposed CERCs would tackle bigger, more complex problems and likely have more diverse research teams, it follows that their budgets would be larger than those of current ERCs. This more complex undertaking could require more resources for supporting roles such as project, cross-project, and cross-center management and/or emerging integrative roles, such as interdisciplinary scientists and data science professionals.29
The funding levels of a number of international center programs are growing and, in some cases, appear to be higher than that of NSF ERCs.30 For example, centers in Singapore, which are modeled on NSF centers, often receive $10 million to $15 million per year.31 The committee makes no recommendation on absolute funding levels but notes that $3 million to $5 million for an ERC in 1985 would translate to between $7 million and $11 million in 2016, accounting for inflation.
In the committee’s view, the alternative model of tackling big problems with a larger number of smaller, more focused centers would be a recipe for higher overhead and transaction costs.
In the absence of a larger appropriation from Congress, the need for larger CERC budgets could be met by reducing the number of centers funded (currently around 20 per year) or by bringing supplementary funding from other federal agencies, international governments, states, the private sector, or foundations. The three examples of possible CERC models discussed in Chapter 7 all involve some form of cost sharing.
27 Note, however, that agencies that fund research via the top-down approach do typically maintain staff technical expertise and receive feedback through the consultative process that program officers have with their research discipline communities.
29 National Cancer Institute, “Team Science Toolkit,” https://www.teamsciencetoolkit.cancer.gov/Public/ExpertBlog.aspx?tid=4&rid=1838, accessed August 28, 2016.
30 E. O’Sullivan, 2016, “A Review of International Approaches to Center-Based, Multidisciplinary Engineering Research,” a commissioned paper for this study, available at https://www.nae.edu/Projects/147474.aspx.
31 Curt Carlson, committee member, personal communication.
The desire for reduced administrative burdens has been frequently expressed in ERC reviews and surveys. Extensive, ongoing data collection for ERCs is needed to fulfill NSF reporting requirements32 (e.g., annual and renewal reports), and this can be difficult when some key information may be held by dispersed team members. Due to privacy concerns, collecting demographic information to demonstrate progress on diversity goals may pose challenges, and considerable effort may be required to capture data reflective of interaction with industry partners. ERCs also must undergo annual site visits, which require considerable planning, and longer site visits when they apply for renewal in years three and six. Additional staff time and administrative work are needed to establish and maintain the various external and internal boards and councils that NSF requires as part of the ERC infrastructure (see Appendix C).
Many international center programs may have lighter annual reporting requirements compared with ERC programs. Although there is significant variation in practice from program to program in terms of reporting on progress, a number of international center directors interviewed as part of this study quickly volunteered that their annual reporting requirements and midterm reviews are not too onerous. It was also suggested by some of those interviewed that management information tools and IT systems were reducing the burden of annual reporting, making it easier to collect and collate journal articles, conference papers, patents, and so on, and to gather information about outreach and impact activities.33
While these concerns are not new, as part of its re-visioning effort, NSF should review its accountability procedures and minimize bureaucratic reporting requirements, with an eye to identifying what outcomes are essential to report, what might be nice to know, and what is unnecessary.
By placing bold bets on a small number of well-funded, prestigious centers focused on engineering solutions to society’s greatest challenges, NSF will create excitement in the engineering community, as well as the natural and social science communities, that will attract the best students, faculty, and industry partners to the CERCs. As they reach for grand technological challenges, CERCs will also build U.S. technological competitiveness and capacity by educating a diverse group of students, staff, and faculty in the rapid translation of research results into products with impact.
While only a small fraction of U.S. engineering graduates will be directly touched by these centers during their university experience, each CERC will have impact far greater than the size of the initial federal investment. CERCs can serve as experimental testbeds to develop engineering curriculum modules, research methods, and work products of durable intellectual value that can be scaled up and disseminated widely, such that the overall impact of the centers is magnified many-fold (see Chapters 3 and 4).
In summary, these new CERCs will
- Address the grand challenges facing society by leveraging the convergence of science, engineering, medical, and—importantly—social science disciplines to accelerate the discovery of new knowledge, create new methods and tools, and develop new products;
32 NSF’s guidance document (2017), intended to help centers prepare annual reports, is 52 pages long. See NSF, 2017, FY 2017 Guidelines for Preparing Annual Reports and Renewal Proposals for the Engineering Research Centers and Nanoscience Engineering Research Centers. Classes of 2006-2015, January, https://www.erc-reports.org/public/download-document?fileName=FY2017_Annual_Reporting_Guidelines.docx; and a separate document (ICF International, 2017) designed to support centers’ use of the agency’s online data system, ERCWeb, is 60 pages long. See ICF International, 2017, Guidelines for ERCWeb Data Entry for the Engineering Research Centers. FY2017, January, prepared for National Science Foundation, Directorate for Engineering, Division of Engineering Education and Centers, https://www.erc-reports.org/public/download-document?fileName=FY2017_ERCWeb_Data_Entry_Guidelines.doc.
33 E. O’Sullivan, 2016, “A Review of International Approaches to Center-Based, Multidisciplinary Engineering Research,” a commissioned paper for this study, available at https://www.nae.edu/Projects/147474.aspx.
- Embrace the best practices of team research and value creation, using advances in information technology, artificial intelligence, social media, and virtual reality to enable deep collaboration that accelerates research advances and innovation in an increasingly interconnected world;
- Leverage the emerging fields of data science and analytics to inform research directions and enhance team research;
- Create new engineering platforms and tools upon which others will build, accelerating the pace of research and innovation;
- Attract the best students, faculty, and industry collaborators, who will accelerate translation and innovation in a dynamic and exciting experiential learning environment;
- Provide students with the full range of skills they need to be leaders in an increasingly interconnected and multidisciplinary world; and
- Develop meaningful domestic and international partnerships with industry, government, nonprofit and philanthropic organizations, and the venture capital community to bring about major advances.
CERC structure and operations (see Chapter 5) will be determined by the goals of the particular initiative and so will not be “one size fits all.” One model might employ a challenge- or prize-based approach to research, with different external teams competing to solve a problem; another might combine resources and talent from federal, state, and local sources to address a regionally important need. Examples of what these alternative models might look like are discussed in Chapter 7.