3

New Directions in University Interactions with Industry and National Laboratories

The second panel session aimed to address NSF Director France Córdova’s question regarding academic-industry partnership models that might effectively promote advances in user-inspired basic and translational research, in addition to considering university-national laboratory partnerships. The three panelists in this session were David Parekh, corporate vice president for research at United Technologies and director of the United Technologies Research Center; Thomas Siebel, founder and chief executive officer of C3 IoT; and Kelly O. Sullivan, manager of institutional science and technology investments at PNNL. A discussion was moderated by Maxine Savitz, retired general manager for technology partnerships at Honeywell, Inc., and retired vice president of the NAE.

AN INDUSTRY VIEW OF UNIVERSITY PARTNERSHIPS

From his perspective of having spent time in academia and industry, and being a product of an ERC, David Parekh thinks of the relationship between industry and academia in terms of a three-dimensional space, with one axis representing how faculty organize themselves—from individuals to clusters across universities; the second axis representing industry, again ranging from the research an individual company engages in to consortium-based research; and the third axis representing funding sources, ranging from government to industry and combinations of the two. Things work well in each of the quadrants of that three-dimensional space, he said, although how the arrangements are structured will vary. Intellectual property, for example, is simpler to structure when one company is funding one investigator than when there are 10 different corporations and universities involved.

Parekh explained that his experience has taught him that there are five characteristics of successful and sustainable collaborations between industry and academia. The first is alignment and focus on a compelling idea. This does not mean that both parties need to have the same goals—an academic center might want to use a novel company-developed technology, and the company may be interested in cultivating and recruiting talented graduates—but they do need to be aligned on strategy, final outcomes, and metrics.

The second characteristic of a successful collaboration is having a compelling idea, something new that takes advantage of the skills and infrastructure that each collaborator can bring to a particular project. The third characteristic is critical mass—the idea that together the collaborators create a capacity that is more than the sum of what each partner alone can muster.

The fourth characteristic Parekh noted is talent and leadership. “We have seen many collaborations that have the right ingredients but a lack of the right leadership causes them to fall apart,” he said. Relationships are the final piece of the puzzle, he added, and the one that sustains the best collaborations. In thinking about ERCs, Parekh said that whatever can be done to promote relationships, to promote engagement among the people involved in a project, is probably the most important factor for ensuring success and sustainability.

Looking ahead, Parekh sees three challenges for centers that can slow progress if not addressed. One is complexity involved in negotiating issues regarding intellectual property, given the multitude of

models that companies and universities have. In his mind, the reinvention of intellectual property agreements that occurs every time a new center is formed is a waste of time and energy that causes unnecessary consternation among the partners. Related to this is the fact that industry and academia move at different clock speeds. Industry, he said, operates with a certain sense of urgency, while academia is more deliberative, and both approaches are valuable. The key is to have in place a mechanism that addresses the inevitable conflicts that occur when patience and impatience clash during the discussions that take place when establishing a center involving academic and industrial partners.

Parekh’s second challenge involves cybersecurity and protecting the information flowing between collaborators while also creating an environment for the free flow of information among partners. The third related challenge is export control. He noted that if research is conducted in Europe, for example, and the data are sent to a collaborator in the United States, the collaborator may then, depending on the project, not be allowed to send the data back to Europe. “The complexity of managing the flow of information is a fundamental barrier to collaboration,” said Parekh.

When thinking about the grand global challenges that previous speakers listed and the technologies being developed today, Parekh sees two overlays that he believes are changing how centers might need to have a different emphasis in the future. One overlay is the need to think in a systems context in a way that integrates that way of thinking into every aspect of how a center operates and approaches a problem. “This is not so much about establishing a multidisciplinary team as it is about the science of integration, about connecting the dots with new models of thought,” said Parekh. The second overlay is the dimension of the person, he explained. “Much of our work is on technology, but the human really matters,” said Parekh. People’s experiences with technology in their lives affects their expectations from companies, which in turn changes how companies conduct business. That human dimension will need to be reflected in how centers are established and operate, he said.

In closing, Parekh called for greater engagement between industry and academia that goes beyond the twice-a-year meeting between researchers and an industry advisory panel. “Industry and academia need to work side by side, and taking away the barriers that prevent that kind of collaboration is important,” said Parekh. He also suggested that future centers should focus on compiling and archiving data for the community at large and on developing tools to interrogate high-quality data. Centers could also serve, he said, as compelling models of how to promote engagement and as open sources of tools developed as part of the centers’ research efforts.

PHILANTHROPY’S ROLE IN FOSTERING HIGH-IMPACT COLLABORATIONS

In his presentation, Thomas Siebel discussed three examples of multidisciplinary collaborative efforts that he thought could provide some useful insights when considering the future of NSF’s ERCs. The first example was the Siebel Scholars, which since 2000 has provided a $35,000 award to the top graduate students selected by the deans from 25 graduate schools at 12 U.S. universities and 4 in Europe and Asia. As of the date of this symposium, the Siebel Scholars community, numbering more than 1,000, has authored 373 patents, founded more than 150 companies, launched more than 1,100 products, founded and operated 57 nonprofits, served on nearly 350 boards of directors, and managed more than $2.7 trillion in assets. In addition, said Siebel, these namesake scholars volunteer for more than 19,000 hours annually.

While these accomplishments are notable, he said the best part of this program is the collaborations that develop among the scholars. Each year, on the occasion of welcoming the new scholars, current and past scholars gather to discuss one topic in depth. For example, the inaugural conference, held in 2000 at the University of Chicago, discussed the threat of nuclear proliferation with guests Alexander Haig, John Major, and Roberts Gates. Other annual gatherings have focused on stem cell research, the U.S. criminal justice system, water supplies, and the economics of alternative energy. The next conference, he noted, will focus on grid cybersecurity.

Out of the gathering on stem cell research, the Siebel Foundation funded the Siebel Stem Cell Institute, a collaboration involving the University of California, Berkeley, the Howard Hughes Medical Institute, and the Stanford Institute for Stem Cell Biology and Regenerative Medicine. This institute now has 83 Siebel investigators from 11 countries and has provided seed grants to support 33 researchers working on 14 innovative collaborative projects, said Siebel.

Discussions among the scholars also led to the founding of the Siebel Energy Institute, an international consortium of universities focused on advancing data analytics research for energy systems, including the smart grid and the oil and gas industry, in an open, collaborative and publicly available manner. One goal of this institute, with initial funding of $10 million, is to use machine learning to dramatically increase the safety and reliability of the power generation grid while lowering costs and reducing its environmental impact. The Siebel Energy Institute has put out calls for papers to advance the science of machine learning in meaningful ways. Awardees are funded to write the papers and then conduct the research that will advance machine learning, with the papers and the research findings all put into the public domain.

COLLABORATING WITH A NATIONAL LABORATORY

The U.S. National Laboratory system comprises 17 institutions that are part of the Department of Energy (DOE) (Figure 3.1). Each laboratory, explained Kelly Sullivan, is owned by the U.S. government, and all but one are operated by contractors. PNNL, her home institution, is operated by Battelle Memorial Institute, and it has four locations in Washington and Oregon, including the only marine research facility in the DOE system. PNNL’s $995 million fiscal year (FY) 2015 operating budget, she explained, was funded by a variety of sources, including DOE, the Department of Homeland Security, other U.S. government agencies, and industry.

As an example of the type of project PNNL scientists and engineers tackle, Sullivan pointed to the laboratory’s work developing algorithms that enable all of the data coming in from a national network of radiation sensors installed at the nation’s ports of entry. At one time, the data from this network took 10 days to analyze, but PNNL’s team developed software that analyzes the data in 10 seconds. “The Department of Homeland Security now has real-time analysis of data from the nation’s ports of entry,” said Sullivan. U.S. allies have also deployed this software, she added.

Sullivan noted that PNNL’s mission is to “transform the world through courageous discovery and innovation.” To fulfill that mission, PNNL operates on a timescale that is somewhat faster than that of a university, creating a challenge for PNNL staff who are interested in working with graduate students on a thesis or dissertation project. “It is doable, but our university partner has to be willing to think about different models,” said Sullivan. Such institutions do exist, and PNNL has developed joint appointment programs with Brown University, University of Washington, Washington State University, University of Utah, and others to share, rather than poach, talent. In some cases, PNNL staff work at the partner university; in other cases, faculty or staff work at PNNL. “Everything they do belongs to both institutions, which made the lawyers very uncomfortable,” said Sullivan, referring to intellectual property and licensing issues.

PNNL, said Sullivan, has funded a number of joint research institutes, some of which succeeded while others failed. One challenge has been that institute researchers consider themselves to be affiliated with the research institute rather than PNNL, which leaves the broader research community wondering what PNNL brings to the table. “We need to figure out ways where the organizations involved in these research centers, including PNNL, all get credit for the work being done.” Sullivan also noted the perverse situation regarding PNNL and the other five national laboratories that Battelle operates for DOE that makes it difficult for her to spend time at one of these other national laboratories. “I would have to change employers, which seems silly but is the truth,” said Sullivan, who added that she is trying to develop some mechanism that supports, rather than inhibits, collaborations and staff exchanges between the national laboratories.

In thinking about the next 10 years, Sullivan suggested that the committee that organized this symposium might consider how employment issues can be addressed to enable the most fruitful collaborations. She also thought that universities should think about how to leverage the expertise and infrastructure at the national laboratories that the nation is already supporting. “We want your students to be working with us,” said Sullivan. “They energize our staff and they help us be more creative, so think about how we can be a part of your campus. We do not give degrees, but certainly have a great deal to teach.”

Referring to the Siebel Energy Institute and its focus on the grid, Sullivan said that PNNL and the National Renewable Energy Laboratory are co-leaders of the Smart Grid Lab Consortium, which she said is working on modernizing the grid and enabling the 3,600 utility operators to work together better. As an example, she noted that although each of the utilities generate data they do not share with one another, they do share data with the consortium under a proprietary agreement to enable research to develop better ways of managing the grid at its maximum capacity. One of the tools the consortium has developed, she said, has enabled PJM, a regional transmission organization that coordinates the movement of wholesale electricity in all or parts of 13 states and the District of Columbia, to save $100 million annually in operating costs.

In closing, Sullivan said, “I just want to make sure people understand that the national laboratories are your asset. They are your resource, so let us figure out how you can take advantage of that system.”

DISCUSSION

Maxine Savitz started the discussion by asking Parekh to comment on how other countries are looking at research centers that might be different from the way the United States approaches them.

Parekh replied that he sees differences in terms of focus—for example, Germany’s focus on advanced manufacturing technologies such as additive manufacturing. He also said that some countries take a more integrated approach that goes from discovery to commercialization, which he said can foster powerful collaborations. He observed that the rest of the world is trying to copy the United States with regard to establishing technology transfer offices; although he also noted that some of these countries are getting ahead of themselves because they are not yet generating a body of intellectual property that needs protecting and licensing.

Andreas Cangellaris asked Siebel to comment on the role that philanthropic organizations can play in energizing innovation as entrepreneurs try to find ways in which their interests can capitalize on the academic ecosystem and federal funding agencies. Siebel replied that the proliferation of multidisciplinary centers for design, which not only include engineers and scientists but experts in the fine arts, performing arts, liberal arts, and humanities, is leading to “amazingly creative work” in product design and ideation. He believes the proliferation of these centers, with funding from the nonprofit world, bodes well for the future of American product design and innovation.

Cangellaris voiced his opinion that the national laboratories need to start talking openly about the impact they have on the nation’s technological development. He called on those organizations that partner with the national laboratories to help get the word out about the enormous value the nation gets from the national laboratories. Sullivan agreed wholeheartedly and added that the Cold War origins of the national laboratories is partly to blame for the low-key approach they have taken to publicity, but that is changing now to some extent, given that the national laboratories now engage in a great deal of unclassified work. Sullivan liked the idea of getting the partners to be more vocal about the role of the national laboratories, noting that it is better when someone else does the bragging.

Responding to a question about funding the riskiest research, and whether that is a place where nonprofit organizations could help, Parekh noted that industry does fund early-stage research through collaborations that address specific fundamental problems that need to be addressed as part of bigger projects. For example, United Technologies is doing cutting-edge research on modeling two-phase flows, but when it needs theoretical work to move the project forward, it turns to partners in academia who are best equipped to do theory development. Siebel then commented that philanthropy is primarily an American phenomenon and that in his opinion, philanthropic organizations that want to make a big impact on important problems should be making big bets on smart people rather than spreading their money around among many projects. Jean-Lou Chameau added that in his experience it is getting harder to fund high-risk projects, and that small, forward-thinking foundations could become ideal partners for such projects.

An unidentified participant asked the panelist for any ideas on how to improve academia’s access to research data compiled by the private sector. Parekh said the key is collaboration, although there are barriers to overcome when it comes to providing access to closely held data and intellectual property. In his experience, although answering important questions might require giving up intellectual property, the payback is usually bigger than the cost. The one place where data exchange faces the fewest barriers is when industry participates in a government-funded program. “What is nice about those programs is that there is no debate about the terms and conditions, which makes it easy to get started. Then you can build the joint interests and commonality that makes things easier later,” said Parekh.

The final question, from another unidentified participant, asked the panelists for their ideas on how the NSF centers can help students develop good integrative skills, given the growing importance of multidisciplinary research. Sullivan responded that systems integration is something PNNL does a great deal of and her experience has been that students do not have a good understanding of how to work in large teams and communicate across disciplines. Those two skills, and safety training, are areas in which she believes the ERCs should put more emphasis. Parekh said he would emphasize the importance of model-based approaches and platform-based design to reduce the number of iterations it takes to refine technologies and improve the transferability of research into practice and products.