The symposium’s first panel session featured three presentations designed to stimulate discussion about engineering research in the broader international context. The three panelists were Arvind Krishna, senior vice president for global research at IBM; Eoin O’Sullivan, director of the Centre for Science, Technology, and Innovation Policy at the University of Cambridge; and Jean-Lou Chameau, president of the King Abdullah University of Science and Technology. An open discussion moderated by David Walt, university professor at Tufts University, followed the presentations.
A common thread connecting some of the world’s most pressing problems, such as how to help people age well, maximize water efficiency, assure food safety, reduce pollution, and manage energy production, is data, vast amounts of data, said Arvind Krishna. High-quality data are a critical resource, he said, and there is a need for new science to extract value from data to create a radically more efficient infrastructure that will enable the United States to continue to be a global leader technologically and economically. As an example, he said a small electrical grid can produce 150 petabytes of data that is just waiting to be analyzed in a way that would almost certainly lead to more efficient operations. Similarly, a single oil well can produce terabytes of data per day. Currently, the vast majority of those data are discarded, simply because the data producers do not know how to use it. Yet by some estimates, said Krishna, the output of existing wells could be improved by as much as 30 percent if operated more efficiently.
Enabling the analytical processes that would produce the type of productivity improvements Krishna envisions requires high-resolution, real-time, open data sets and new methods of getting shared value out of closed data sets, such as anonymized consolidated electronic health records, seismic exploration data, and the results of drug development efforts. One role for the next-generation ERCs, said Krishna, should be to generate open data sets and methods for mining closed data sets in a way that protects intellectual property rights. Another focus should be on creating new information theory for making mission-critical decisions with data using high-performance computing. He noted that while NSF has played an important role in developing high-performance computing, the entire computer industry, including the processor and semiconductor segments, are having an impact in this area—and are in turn impacted by it—at a scale that is “vastly underestimated by the average individuals.”
Cognitive computing is making important strides, with important contributions from industry and academia, said Krishna, but further advances will require the development of new science in both hardware and software development. Krishna believes, as an example, that more work is needed to measure the entropy of data sets to determine the quality, richness, and diversity of a data set and whether it is complete enough to enable cognitive computing applications, such as anomaly detection, with context and behavioral analysis that ensures privacy. New science could also enable applications in the areas of risk-based capacity planning for infrastructure development, control of viral propagation, and computational creativity.
While new processor technologies using field-programmable gate arrays, for example, and advanced graphical processing units are producing multiple-fold improvements over traditional computing with regard to machine learning, those kind of incremental advances will not change the game, said Krishna. Neuromorphic computing, inspired by the architecture and operations of the brain, has already improved processing time on selected video tasks by three orders of magnitude, he said, but highly speculative technology still being investigated could increase the speed for training neural networks by four to five orders of magnitude. The timeline for turning these proof-of-concept demonstrations into implemented technologies will be longer than most companies can support, however, so this is a place where NSF could play a role through its centers by bringing researchers from academia and industry together to shorten that timeline. “The nation stands to benefit on the order of trillions of dollars if we can do that,” said Krishna.
A number of global megatrends are shaping national economies and their innovation systems, said Eoin O’Sullivan, and these include threats to global stability, digitalization, globalization, demographic changes, changing consumer habits, accelerating product life cycles, external industrial policy trends, urbanization, and a focus on sustainability. While the precise list of innovation policy responses and emphases will change depending on a country’s industrial strengths and specific challenges, there is a consensus among those who think about such issues on how policy might respond. These responses will highlight a sense of urgency and draw more attention to challenge-led innovation. As a consequence, said O’Sullivan, there will be a greater emphasis on engaging a broader range of users who have something to say about these far-ranging societal problems. Innovation policy will focus on steps to increase the speed of translating innovation into the marketplace, not just with respect to promoting the development of new prototypes and establishment of test beds, but also addressing scaling and manufacturability. Policy is also expected to stress global connectedness of innovation centers, yet also reflect the tension between collaboration and the need to capture value in a way that ensures an appropriate net flow of knowledge and innovation to the home country.
One change for centers that Sullivan believes will be important will be an increasing emphasis for them to fit into a broader system context, to connect to other innovation actors, such as the national research laboratories, applied research and development (R&D) institutes, and national metrology laboratories. Innovation policy, he predicted, will more easily enable linkages between these different components of an innovation ecosystem, which in turn will attract R&D funds at a time when such funds are becoming globally mobile.1 Today, he noted, U.S.-headquartered companies are spending almost as much on R&D outside the country as inside the United States.2 Recently, he said, Asia has become the number one destination for investment in corporate R&D.
The sense of urgency from competition that nations are feeling is reflected in recent policy documents of many national governments around the world referring to center-based programs, said O’Sullivan. While this may always be the case for countries such as Ireland and Singapore that rely heavily on R&D as a core component of national economic policy, this trend is new for nations such as Germany and the United Kingdom. These policy documents are not only stressing the importance of research centers as sources of innovation, but also their role as honest brokers in helping to develop supply chains and arrange for interactions and collaborations among larger firms. Research center administrators and staff, said O’Sullivan, are now trying to understand how they need to create and
1 B. Jaruzelski, K. Schwartz, and S. Volker, 2015, Innovation’s new world order, Strategy+Business, Issue 81 (Winter), October 27.
2 B. Jaruzelski, K. Schwartz, and V. Staack, “Where Companies Spend Their R&D Money,” [interactive], Strategy+Business, Issue 81 (Winter), October 27, 2015
nurture an environment that creates a visible critical mass of expertise to draw in corporate investment in an increasingly competitive global environment.
At the same time that countries talk more about national value capture, their innovation policies are emphasizing the need for their centers to be connected globally to other centers. There is an interesting ongoing debate, said O’Sullivan, about net knowledge flow at a time when nations with mature innovation economies are partnering with countries whose innovation economies are less mature. This division of labor tends to have engineering and scientific research take place in the more mature nations while prototyping, scale-up, and manufacturing occur in what he called the “catch-up” economies.
Another trend O’Sullivan has noticed has been the increasing number of thematic calls for center proposals rather than more general research calls. As a result, there are an increasing number of national workshops and road-mapping strategies that aim to create new communities and identify networks among individual potential collaborators. These workshops and mapping exercises tend to highlight key barriers to address and prioritize innovation challenges in such a way that the resulting proposals are often better and more thoughtfully crafted. O’Sullivan said he has also seen an increasing emphasis on manufacturability, piloting, and scale-up in research center goals, which has translated into more funding for such activities at the center level and an increasing emphasis on being agile and responsive to market opportunity. In Germany, for example, research center programs are encouraged to transfer to another research institute or center if it offers the opportunity to take advantage of new commercial opportunities.
These trends, said O’Sullivan, have led to centers being more engaged with business schools, economics departments, and experts in operations management. They are also challenging applicants for center programs to state in their proposals how they might engage other partners and leverage the expertise and infrastructure at other innovation institutes to maximize the impact of their work on the high-level global megatrends affecting national economies. This change is occurring even in countries such as Germany, which has its well-established Fraunhofer Institutes that work with industry and universities to scale cutting-edge research into real working technologies on an industrial timetable. Germany’s new Forschungscampus Initiative creates center-like endeavors on university campuses to bring together university, government, and industry researchers in partnerships to address challenges that no one organization can tackle.
Over the past 15 years, said Jean-Lou Chameau, faculty and students are become more mobile, more global, and more opportunistic. Successful senior faculty, he added, have become particularly global, often having their main laboratory on one campus and auxiliary laboratories spread across the country and the world. These laboratories are then networked, creating an interactive, multicultural ecosystem for innovation. The challenge that results from this type of arrangement is for university structure and policy to keep up with these new models that faculty are creating in order to nurture allegiance to the home university. “This is the type of situation that can test the skills of deans and provosts,” said Chameau.
At the same time, nations in Asia, Europe, and the Middle East have recognized this trend in faculty mobility and are creating opportunities at research institutes to attract senior faculty and the brightest students. Korea, Singapore, and Switzerland have been particularly adept at developing this type of center, he noted. “The intent is to entice researchers and students with a strong focus and long-term sustained support,” said Chameau. “This issue of long-term sustained support is something to think about with regard to the U.S. engineering research centers.” In Chameau’s opinion, there is an opportunity for U.S. agencies and research universities to consider joint leadership with foreign institutions that are currently driving the creation of such centers. “Universities could integrate their campuses into research networks by building upon the aspirations of faculty and students who are already doing this on their own,” he explained.
He also reiterated Krishna’s and O’Sullivan’s statements that corporations, particularly in Asia, the Middle East, and Europe, have become active participants in, and even drivers of, the creation of these new research centers, even more than they may have been in the past with the ERCs. Today, companies are becoming involved not only in defining the research scope of these centers, said Chameau, but are also bringing new partners to the table and participating in the research mission from the very inception of these centers. One positive result of the private sector’s increasing role in center formation, he added, is that there has been an improvement in resolving conflicts of interest and developing commercialization and disclosure policies. In addition, the centers have been given more of a mandate to play a strong role in knowledge transfer and economic development, so much so that the goal of driving innovations and technology transfer have become the vision of these centers and essential components of their design. This is much different from the way universities have been structured in the past. An important aspect of this mandate that Chameau sees playing out in the United States, as well as in his current location in Saudi Arabia, is that centers are actively nurturing and leveraging the entrepreneurial spirit and innovative mindset of students.
International partnerships, with funding from both industry and national research agencies, are becoming more prominent in the newest generation of centers, said Chameau. In Europe, this has created a more reassuring environment for faculty with regard to sustainable funding, which in turn is enabling these centers to acquire the best equipment and attract better-quality researchers. Yet even with the creation of these new centers and more sustainable funding streams, the United States, in his opinion, is still regarded as the place that researchers want to be because of the amazing track record of U.S. institutions and agencies, including NSF. As a result, the United States has a great opportunity to focus on the grand challenges that O’Sullivan listed, which transcend national boundaries, and to be the leader in promoting joint investment in research and human capital worldwide. “I think the NSF could take a strong leadership role and get the benefits, including financial support and resource sharing,” said Chameau.
Going forward, Chameau believes that just as Facebook is described as social by design, an increasing number of centers will be global by design. He also predicts that universities will become organized increasingly around research centers rather than being driven by academic units, which is now the case, and become more mission-oriented toward solving the great issues facing society. At the same time, he cautioned, universities and research centers will have to recognize the need for a mix of mission- and goal-oriented research and curiosity-driven research. It is important, he said, to not lose the aspects of the U.S. research enterprise that have made it so successful at both producing the fundamental knowledge that serves as the foundation for innovation and harnessing the entrepreneurial drive that turns knowledge into economic activity. One area in which the United States could do better, however, is in preparing postdocs for life in industry.
In closing, Chameau said that he is a strong believer that the success of the United States in research and innovation comes from tapping into the creativity of the nation’s talent and enabling its scientists and engineers to pursue research without a specific product immediately in mind. “Although I told you that I love very much the mission-oriented goal, I think it has to be done in a way that it enables curiosity-driven work,” said Chameau. “At the same time, people should be encouraged to take the steps to translation and commercialization. It should be part of their expectation.”
Keith Roper, from NSF, asked the panelists for their ideas on how to balance global by design with the national boundaries associated with finance and intellectual property. Chameau replied that what he means by global by design is a structure that brings together a group of researchers from different parts of the world who want to pursue an idea and who believe strongly that they will be more successful as a group than as independent players or even working in independent centers. From his experience, he believes that if you have that kind of commitment among the group members, the issues of intellectual
property can be resolved, although not always perfectly. O’Sullivan said that increased attention to the global grand challenges he listed is creating more openness to global collaboration because they are of a scale and complexity that no individual country can take them on by themselves. “There will be complementary expertise and capabilities that people will want to reach out to, and as a result there will be a greater scope for a win-win situation in taking on some of these challenges,” said O’Sullivan. As a result, he believes it will become easier to align funding mechanisms between countries. He added that his experience in speaking with large R&D-intensive companies about their major partnerships with universities is that they want to better align and connect their investments at different universities and centers around the world. Krishna added that IBM’s experience working in 195 countries shows that intellectual property issues are not that hard to solve from a global perspective. What is more difficult to address are data issues because data transfer is often controlled by national laws and regulations.
In response to a question about how funding agencies can help the research community develop ideas for new centers designed around specific goals, Chameau said that NSF has done a good job providing small amounts of seed money on an opportunistic basis for exploring certain ideas that could lead to something bigger. Krishna added that incentives and measurements are important, and that incentives are not necessarily about money, although he did not elaborate further.
Kelly Sullivan, from the Pacific Northwest National Laboratory (PNNL), asked the panelists for their ideas on how to deal with the different incentives, reward structures, and missions of various partners involved in a center. Krishna replied that the key is to acknowledge those differences upfront. Andreas Cangellaris, from the University of Illinois, Urbana-Champaign, noted that the challenge universities face today to retain faculty in “hot” areas of research while still sharing their expertise with industry in a way that is creative and sustainable. Krishna responded that Cangellaris had poached four members of his research team over the past 2 years as a way of illustrating the bidirectional nature of this challenge. “I think it is about creating shared and bidirectional systems and allowing people to move back and forth, albeit for limited amounts of time,” said Krishna.